Rapid Woodfuel Assessment 2017 Baseline for

Bidi Bidi Settlement, Uganda

Woodfuel supply/demand and screening of

scenarios to improve energy access and reduce

environmental degradation

May 2017

iii

CONTENTS

ACKNOWLEDGEMENTS IV

ACRONYMS V

EXECUTIVE SUMMARY VI

1. INTRODUCTION 1

1.1 BACKGROUND 1

1.2 OBJECTIVES OF THE STUDY 2

1.3 METHODOLOGY 3

1.3.1 AREA OF INTEREST 3

1.3.2 WOOD ENERGY DATA COLLECTION AND ANALYSIS 4

1.3.3 BIOPHYSICAL FIELD INVENTORY 5

1.3.4 REMOTE SENSING ANALYSIS 8

1.3.5 DEFINITIONS 11

2. WOODFUEL DEMAND RESULTS 13

2.1 POPULATION AND HOUSEHOLD INFORMATION 13

2.2 WOODFUEL CONSUMPTION FOR COOKING AND HEATING 14

2.3 TECHNOLOGIES AND PRACTICES FOR COOKING 17

2.4 CHALLENGES AND RISKS TO ACCESS WOODFUEL 22

3. WOODFUEL SUPPLY RESULTS 24

3.1 RESULTS FROM THE FIELD MEASUREMENTS 24

3.2 MAPPING THE DISTRIBUTION OF LULC 25

3.3 MAPPING CHANGES IN LULC 27

3.4 BIOMASS MAPPING 27

4. POSSIBLE SCENARIOS AND RECOMMENDATIONS 30

4.1 POSSIBLE SCENARIOS 30

4.2 RECOMMENDATIONS 32

5. CONCLUSIONS 35

6. REFERENCES AND FURTHER READING 36

7. ANNEX 37

iv

Acknowledgements

This report was developed by Arturo Gianvenuti, Antonia Ortmann, Eva Kintu and Remi D’Annunzio.

The activities were carried out thanks to the UNHCR financial contribution to the FAO Strategic

Programme 5: Increase the resilience of livelihoods to threats and crises, in the framework of the FAO-

UNHCR “Collaborative partnership for the assessment of woodfuel supply/demand and screening of

improved scenarios to provide support to energy needs of refugees in Bidi Bidi settlement (Uganda)

and reduce forest degradation and deforestation”.

Contributors

Andrea Dekrout, Amare Gebre Egziabher, George William Ilebo, Acacio Jafar Juliao, Paul Quigley,

Agnes Schneidt and Wilfred Wadada (UNHCR); John Begumana Ayongyera, Andrea Borlizzi, Massimo

Castiello, Sergio Innocente, Inge Jonckheere, Damiano Luchetti, Darlene Lutalo, Michela Mancurti,

Giulia Miur, Teopista Nakalema, Mats Nordberg, Simona Sorrenti, Andreas Vollrath and Xia Zuzhang

(FAO); Baker Andreonzi, Charles Ariani, Edwin Beinomugisha, Zabibu Ocogoru and Edward Ssenyonjo

(NFA); Catharine Watson (ICRAF).

v

Acronyms

AGB

AOI

DBH

FAO

Aboveground biomass

Area of interest

Diameter at Breast Height

Food and Agriculture Organization of the United Nations

FGDs Focus Group Discussions

FRA

HH

Global Forest Resources Assessment

Household

IMAD Iteratively Multivariate Alteration Detection

IRC International Rescue Committee

KIs Key informant interviews

LULC

NBS

NFA

Land use/land cover

National Biomass Study

National Forest Authority of Uganda

NGO Non Governmental Organization

OPM Office of the Prime Minister

REDD+ Reducing Emissions from Deforestation and Degradation

SAFE Safe Access to Fuel and Energy

SEPAL

SPGS

System for Earth Observation Data Acquisition Processing and

Analysis for Land Monitoring

Sawlog Production Grant Scheme

UBOS Uganda Bureau of Statistics

UNHCR United Nations High Commissioner for Refugees

VGGT

VHR

Voluntary Guidelines for Governance of Tenure

Very High Resolution

vi

Executive summary

Uganda is currently host to over one million refugees who have fled from famine, conflict and

insecurity in the neighbouring countries of South Sudan, Burundi and the Democratic Republic of the

Congo (DRC). The recent influx of refugees from South Sudan has prompted one of the most severe

humanitarian emergencies in Uganda and led to the establishment of the Bidi Bidi settlement in

Yumbe District in August 2016. The Bidi Bidi settlement is now the largest refugee-hosting area in the

world with 272 206 refugees settled on approximately 250 km2 of land over a total assigned area of

798 km2. Bidi Bidi refugee settlement constitutes more than half of the population of the host district

Yumbe (484 822 people) (UBOS, 2014). This has increased pressure on the environment due to the

felling of trees for setting up the settlement and for meeting household woodfuel demand for cooking

and heating. As such, a rapid woodfuel assessment was jointly initiated by UNHCR and FAO in March

2017 to assess the supply and demand of woodfuel resources in the area. The methodology included

three phases: 1) an assessment of the woodfuel demand, 2) an assessment of the woodfuel supply,

and 3) the identification of relevant inter-linkages, gaps, opportunities and alternative scenarios. Data

and information were obtained through a combination of desk review of existing documents, field

surveys and remote sensing analysis.

Main findings

- Fuelwood consumption. The findings presented in the woodfuel demand part of this study

highlight an estimated average fuelwood consumption for cooking/heating in Bidi Bidi

settlement of 20.5 kg of air-dried wood per household, or 3.5 kg per person per day. For the

purposes of this study, the daily fuelwood consumption is assumed to remain constant over

the year with no seasonality variations. Although in this case only a minor share is used for

heating, an average daily fuelwood consumption of 3.5 kg per day is considered quite high if

compared to similar situations. For instance, in literature the fuelwood consumption for

cooking alone ranges from 0.7 kg to 3 kg per person per day (Gunning, 2014).

For 272 206 refugees registered in Bidi Bidi settlement, according to the results of the sample

household survey, the rate of fuelwood consumption for the total population of the

settlement is about 952 tonnes per day which corresponds approximately at 347 480 tonnes

of fuelwood per year. Since the camp was opened over approximately seven months ago, the

fuelwood consumption is estimated to be 199 920 tonnes thus far. The food rations of

refugees include mainly maize and beans that require a long time for cooking. The household

survey shows moreover that very few (3%) use charcoal for cooking and the charcoal

consumption since the establishment of the settlement is estimated at 672 tonnes, which

corresponds to 3 360 tonnes of fuelwood.

- Mixed use of traditional and fuel-efficient stoves. It was also observed that many refugees

use wet/green wood which increases the volume of fuelwood consumed. Although a higher

proportion of households (56%) use the traditional 3-stone fire for cooking, the survey

results show that the use of fuel-efficient stoves is not new to the refugees in Bidi Bidi and 43%

of refugee households use an improved mud stove and have knowledge and skills on its

vii

construction. The host community seems to have a lower adoption of fuel-efficient stoves

than refugees.

- Heavy burden on women. Another key finding is that the burden of fuelwood collection in

Bidi Bidi settlement is entirely on women. Women make an average of four trips in one week

to collect fuelwood and experience a number of constraints. They carry heavy loads of green

wood which sometimes they use before it dries. Women and girls are exposed to multiple

dangers as they walk long distances for fetching fuelwood including attacks by wild animals,

assault and rape by host communities as well as refugees. The shared use of natural resources

overall has an impact on social relations, causing tension and conflicts between host and

refugee communities.

- Woody biomass also sustains other uses. Woody biomass has been cut down by the refugees

also for building shelters and other household constructions. According to the field

measurements carried out in the field, approx. 0.9 m3 of wood poles were used by each

household for construction. Therefore, the estimated woody biomass used for this purpose

by the 101 500 households estimated by the OPM in Bidi Bidi settlement is 91 350 m3 that

corresponds to 63 945 tonnes (assuming an average wood density of 700 kg/m3).

- Total aboveground biomass stock, growth and land cover/land use assessment. The total

aboveground biomass stock is estimated at 734 614 tonnes (air-dried) within the settlement

boundary and at 1 617 953 tonnes including a buffer area of 5 km. Most aboveground

biomass can be found in open woodlands, but also closed woodlands contribute significant

amounts despite their limited coverage in terms of area. The combined use of field

measurements, land use descriptors and remote sensing analysis of radar and optical data

enabled to map the land cover features within the Bidi Bidi settlement boundary and provided

estimates of biomass for each land cover class.

Within the Bidi Bidi settlement area, the closed woodland feature provides the highest level

of biomass density (191 130 tonnes dry matter over an area of 3 356 ha, representing 26% of

the total woody biomass resources available in the settlement). These resources are

concentrated on a small area – at the edge of the settlement in the north-east – which exposes

them to high pressure for use as woodfuel). Conversely, the amount of biomass that could be

found in the open woodland spreading across an area 10 times as big (39 492 ha) represents

about 50% of the total woody biomass available (364 513 tonnes) and has been identified as

the main potential source for fuelwood collection within Bidi Bidi settlement. It should also

be noted that patches of land assigned for agriculture in the settlement contain significant

amounts of biomass resources (18%).

Because of the strong seasonality effects and agricultural practices (burning), the change in

land cover and land use over this short period of time did not show any significant trend in

woody biomass resources change other than that caused by the establishment of the

settlement itself. However, according to the results of woodfuel demand assessment, an

estimated 203 280 tonnes of fuelwood (including charcoal expressed in fuelwood basis) has

been consumed since the establishment of the settlement and this should be compared to

the estimated current 734 614 tonnes of aboveground biomass stock and the annual

aboveground biomass growth of about 33 300 tonnes per year in the settlement area. If the

viii

5 km buffer area around the settlement boundaries is taken into account, then the estimated

total aboveground biomass stock is 1 588 961 tonnes. It is important to note, however, that

thee area outside the settlement boundaries is also subject to woodfuel demand from the

surrounding host communities.

- Scenarios. By integrating the results from woodfuel demand and supply assessment, the

following future scenarios are projected: 1) Scenario Business As Usual (BAU); 2) Scenario with

improved cookstoves; and 3) Scenario with tree plantations and improved stoves.

(1) Scenario Business as Usual. The woody biomass resources within Bidi Bidi settlement are

scarce and under documented pressure and, with annual fuelwood consumption at 347 480

tonnes, could supply the refugees for only 3 years. This would however be at the cost of a

full depletion of the area in terms of aboveground biomass, representing an issue that needs

to be addressed as soon as possible as refugees already compete with the local communities

for the scarce resources outside the settlement area. Continuous monitoring of these

resources will be needed in the coming months to guide decision making on measures to be

taken to mitigate the pressure.

(2) Scenario with improved cookstoves. The utilization of improved and more energy efficient

cookstoves over of the traditional 3-stone fire can bring fuelwood savings of about 30%

compared to the BAU Scenario. The total fuelwood consumption in Bidi Bidi settlement

would be then about 243 000 tonnes per year instead of 347 480 of BAU Scenario. Under this

Scenario with improved cookstoves, women will also spend less time collecting fuelwood

with more time available for other activities. As the current aboveground biomass growth is

estimated at 33,300 tonnes per year within the settlement area, the annual deficit of woody

biomass growth under this scenario would be 209 700 tonnes. Therefore, according to the

estimated total aboveground biomass stock and annual growth from trees and shrubs within

the settlement area, the total woody biomass resources would supply the current refugees

in Bidi Bidi for 4 years in this scenario.

(3) Scenario with tree plantations and improved cookstoves. To compensate for the annual

biomass loss in the settlement area as result of woodfuel demand, assuming 100% of the

refugee households were to use improved mud stoves (Scenario 2), this scenario includes the

establishment of tree plantations to increase the woody biomass and address the energy

needs for cooking and heating of the refugees. Under this scenario, it is estimated that

between 9 368 and 12 178 hectares would need to be planted with fast growing species

that corresponds to 12-15% of the total land area of Bidi Bidi settlement (798 km2 within

the settlement boundary). Each household would need a minimum area of 50 m x 50 m,

dedicated exclusively to growing timber for fuelwood.

1

1. Introduction

1.1 Background

The drivers and impacts of migration and displacement in acute and protracted crisis are intimately

linked to FAO’s global goals of fighting hunger and achieving food security, reducing rural poverty and

promoting the sustainable use of natural resources including forests and other woodlands. Through

this study and its broader partnership with UNHCR, FAO is addressing its Strategic Objective 5

“Increase the resilience of livelihoods to threats and crises”. This collaboration between FAO and

UNHCR, moreover, is a step towards strengthening UN inter-agency support to displaced people and

to improve the management and sustainability of natural resources in these contexts.

FAO and UNHCR are also working in partnership in the Safe Access to Fuel and Energy (SAFE)

framework to identify and implement appropriate solutions to address the challenges faced by crisis-

affected populations, including both displaced and host communities. The SAFE work adopts a holistic

approach which takes into account the mutually reinforcing linkages between energy access and

associated challenges related to nutrition, forests, climate change, health, gender, protection and

livelihoods.

Uganda is currently hosting over a million refugees who have sought protection from famine, conflict

and insecurity in the neighbouring countries of South Sudan, Burundi and the Democratic Republic of

the Congo. Refugees from South Sudan constitute the most severe humanitarian emergency situations

in Uganda. The recent increasing refugee influx from South Sudan led to the establishment of the Bidi

Bidi settlement in Yumbe District, settled in August 2016. Bidi Bidi settlement is the largest refugee-

hosting areas in the world with 272,206 refugees (OPM – Level 1 Registration Statistics, as of 19

December 2016).

Bidi Bidi refugee settlement constitutes more than half the population of the host District of Yumbe

(484,822 people) (UBOS, 2014). This implies an increased pressure on the environment due to the

cutting of trees for setting up the settlement and for household woodfuel consumption.The refugee

management policy in Uganda is characterised by a non-camp, settlement approach, which is widely

regarded as a best practice example for the region. The 2006 Refugee Act and the 2010 Refugee

Regulations grant prima facie refugee status to all South Sudanese and Congolese and facilitate a food

security and emergency agricultural livelihoods strategy, which allows refugees a plot of land for

agricultural use. In addition, the Uganda Refugee policy promotes the integration of life-saving social

services into national government systems. Specifically, the refugee settlements are integrated within

the host communities, granting refugees:

access to the same services (e.g. health facilities) as nationals;

the right to livelihoods and to establish businesses;

the right to go to school;

freedom of movement;

access to allocated land for agricultural use.

2

These rights constitute a significant opportunity to protect refugees responding to emergency crises

as well as ensuring the conditions for integrating emergency livelihoods within refugee hosting areas

and local government systems, in addition to building self-reliance and resilience.

Within this context, woodfuel is the major source of rural energy in the country and the consumption

and production of fuelwood is estimated to be around 42.4 million m3 (in 2015) and 1.06 million

tonnes of wood charcoal (FAOSTAT, 2015). About 95% of this is consumed as fuelwood. Uganda faces

significant pressure on its forests and woodlands and suffers from a high rate of deforestation and

land degradation. Drastic changes in forest cover have taken place in Uganda during the past few

decades. According to the Global Forest Resources Assessments (FRA) (FAO, 2015), as much as 2.6

million hectares of forest cover have been lost in a period of just 25 years, which is approximatively

13% of the total land area of Uganda. From 1990 to 2000, the average yearly loss of forest cover

amounted to 88,000 hectares. This annual loss amassed to 111 000 hectares per year between 2000

and 2010, and in the following five years this figure grew to a loss of forest cover amounting to 185 000

hectares per year (FAO, 2015).

Fuelwood is a primary source of energy for refugees in Bidi Bidi settlement; charcoal is used by those

households that can afford it. Although National Forest Authority (NFA) has earmarked trees that should

not be cut, all the other trees within the settlement are being cut down due to the sudden increase of

population in this area and the high woodfuel demand for cooking. The energy need for cooking of

large numbers of refugees in Uganda exacerbates already existing fuelwood overharvesting problem

in the country and increases the high risk of rapid degradation of the surrounding environment

including: soils, forests and other woodlands.

With the high number of people crossing the border from South Sudan to Uganda, there is an urgent

to develop supportive sustainable energy access and forest resources management strategies

targeting both the refugee population of Bidi Bidi settlement and the surrounding host community.

1.2 Objectives of the study

The objectives of the study include:

1. Assess the woodfuel demand (fuelwood and charcoal) for cooking and heating and challenges

related to the collection and use of woodfuel for the refugee population of Bidi Bidi settlement.

2. Identify current technologies/practices in use for cooking in the area of interest and the

associated opportunities and challenges.

3. Assess the potential woodfuel supply in the area of interest: state and changes of

aboveground biomass stock, growth and land cover/land use.

4. Provide recommendations for planning interventions to improve energy access, promote

sustainable forest management and contribute in building resilience of affected populations

in Bidi Bidi settlement.

3

1.3 Methodology

This rapid woodfuel assessment builds on the methodology developed in the joint FAO-UNHCR

Technical Handbook “Assessing Woodfuel Supply and Demand in Displacement Settings” (FAO-UNHCR,

2016). The methodology includes three programmatic phases: 1) to assess the woodfuel (fuelwood

and charcoal) demand over the given period of time since the establishment of the settlement

(assuming woodfuel demand is equal to the ongoing woodfuel consumption) and the associated

challenges experienced by the refugees and host community, 2) to assess the woodfuel supply

including the aboveground biomass stock, land cover classification and changes over the period of

time since the establishment of the settlement, and 3) to identify relevant inter-linkages and gaps as

well as to develop possible future scenarios in which improved cookstoves and afforestation

interventions are introduced.

The study involved a combination of desk review of existing documents, field surveys and very high

resolution satellite imagery analysis. Field surveys included the assessment of woodfuel demand in

the refugee settlement and the assessment of the biophysical parameters of forest and other

woodlands. Satellite imagery analysis was performed combining semi-automatic classification of the

land with change detection processes.

1.3.1 Area of interest

The area of interest (AOI, Figure 1) is composed of the area within the settlement boundary and a

surround buffer area of 5 km (estimated as possible area for fuelwood collection outside the

settlement). The settlement boundary comprises an area of 798 km2 within which five zones, covering

a total of 250 km2, were established. The 5 km buffer around the settlement boundary would bring

the total area to 1 617 km2.

4

Figure 1: Extent of the area of interest, with the five zones defined by UNHCR and the 5 km overall buffer area around the settlement

1.3.2 Wood energy data collection and analysis

The woodfuel demand assessment was carried out through household quantitative surveys and

qualitative interviews and focus group discussions in the refugee and host communities. The objective

was to provide data and information on fuelwood and charcoal consumption, the average time spent

to collect fuelwood by the household, the types of cooking systems used for cooking and the main

challenges. The process was of a participatory nature with discussions in the field at different stages

between FAO and UNHCR staff; key informant interviews were held with the Office of the Prime

Minister (OPM), District Officials and a team of International Rescue Committee (IRC) Community

Workers to gather specific contextual and complementary information relating to a broader

understanding of environmental and energy factors in Bidi Bidi settlement. The wood energy and

sociocultural data and information were collected by a FAO and UNHCR team with the support of four

enumerators from IRC between 28 February and 3 March 2017. The sampling design for the household

survey involved: i) the selection of the zones within the settlement to be surveyed; ii) an estimate of

the sample size according to the timeframe and resources available; iii) random selection of the

sample.

Four data collectors who are community workers in the settlement were given a quick orientation of

the household survey after which they tested the tools with some initial households under the

supervision of FAO and UNHCR team. Households were then selected randomly in the different parts

of the selected zones within Bidi Bidi settlement; respondents were mainly women, since they had

the knowledge on fuelwood collection and use in the households or the head of households. Data

5

collection was carried out in three zones (1, 3, 4) and a total of 72 households were interviewed to

collect quantitative and qualitative data through questionnaire and direct interviews.

In order to quantify the daily woodfuel (fuelwood and charcoal) consumption, the assessment team

weighed the estimated amount of woodfuel for cooking and heating using a digital hanging scale and

in the case of fuelwood they also took note on whether it was wet/green or dry based on a visual

observation of the wood. Further observations were also made and photographs taken of current fuel-

efficient devices that were being used in households. This data was analysed to estimate the average

daily wood consumption per person and this result was scaled up to the settlement level.

Focus Group Discussions (FGDs) were undertaken in Bidi Bidi settlement as well as in one village of

the host community. Group discussions were conducted with mixed age, social groups and gender.

Key informant interviews (KIs) were held in the settlement as well as with authorities at District level.

During data collection some substantive checking with data collectors was done while still in the

settlement to verify some occurrences. The data from the household survey was then entered and

analysed using Excel sheets for occurrence and proportions as well as cross tabulation of parameters.

The data was compiled in statistics tables and graphics. The data analysis was then supported by the

application of the SAFE Tool – Woodfuel Assessment in Displacement Settings to identify the baseline

situation and to simulate alternative scenarios in which improved cookstoves and tree plantations are

introduced to increase energy efficiency for cooking, reduce energy consumption and increase woody

biomass growth. The alternative possible scenarios analysed in this study include 1) Business as Usual

(BAU) Scenario, 2) Scenario with improved cookstoves to reduce woodfuel demand, and 3) Scenario

with new tree plantations and improved stoves.

1.3.3 Biophysical field inventory

Biophysical field data was used to estimate stocks of fuelwood for different land use/land cover (LULC)

classes. These LULC classes were mapped through remote sensing which allows mapping the

distribution of fuelwood resources and assessing stock changes before and after the establishment of

the refugee settlement.

Sampling design

For the sampling design, a systematic grid with a spacing of 2 km x 2 km was generated over the

settlement area. A first visual interpretation of the grid points was performed to establish a

preliminary stratification of the land cover in the area. This was based on a combination of Sentinel-2

imagery for the beginning of 2017 and freely available very high resolution (VHR) imagery for the time

before the establishment of the settlement.

Data from the National Forest Inventory were then used against the classes found in the AOI to

determine the number of field plots to measure the biomass content of the relevant land cover classes.

Based on the stratification of the relevant classes (Table 5 in the results section), a total of 32 points

were randomly drawn within the settlement boundaries, to be further measured in the field (Figure

2).

6

Figure 2: Systematic grid with a spacing of 2 km x 2 km and the 32 final plot locations

Plot design and field data collection

For each of those field points, a circular plot with a radius of 12.6 m was chosen, resulting in an

approximate size of 500 m2. Within the plot, several subplots were measured (Figure 3) so as to follow

the methodology used by the National Forest Inventory.

Within the first quadrant of the plot, all shrubs were measured including base diameter, crown and

average height. For all standing trees (live and dead) of 5 cm diameter and above height, diameter at

breast height (DBH) and tree species were recorded. In the remaining subplots, standing trees with a

diameter of 10 cm and above were measured.

In the plot center, within a radius of 4 m, all saplings and deadwood were measured. Deadwood was

collected and weighed when possible. If the deadwood was too heavy to be measured, but its length

and diameter were recorded.

7

Figure 3: Plot design for field inventory

In addition to these dendrometric measurements, the following land cover/land use descriptors were

collected for each plot:

- Slope, orientation, altitude and accessibility

- Major land use and land cover types

- Proportioning of each land cover/land use

- Fire evidence

- Grazing intensity

- Undergrowth type

The biophysical data was collected by specialists of the Inventory team of the National Forestry

Authority of Uganda (NFA) between 28 February and 11 March 2017. The data was recorded using

Open Foris Collect Mobile, an Android App that facilitates fast, intuitive and flexible data collection for

field-based surveys (http://www.openforis.org/tools/collect-mobile.html) (Figure 4).

8

Figure 4: Collect Mobile being used for field inventory in Uganda (Uganda National REDD+ Secretariat, MWE)

Estimating biomass stocks and annual increment

Above ground biomass for each plot was calculated from field data using Chave et al. (2014) allometric

equations which are also used for the National Biomass Study (NBS) in Uganda. R scripts already

developed for the NBS for REDD+ were used to estimate stocks. For the final estimates, plot level

results were aggregated by land use/land cover (LULC) class which had been assigned to each plot

during the field inventory. Biomass estimates by LULC class are calculated as mean, with 95%

confidence intervals.

Biomass annual increment were taken from NBS which provides them as average for different LULC

classes for the different agroecological zones of Uganda (Forest Department 2002). Yumbe District

falls into the semi-moist lowland zone.

1.3.4 Remote sensing analysis

The remote sensing analysis aimed at mapping LULC before and after the establishment of the

settlement in order to estimate changes in biomass. LULC was mapped for the whole settlement and

for an additional buffer of 5 km around the settlement boundary. This buffer was applied in order to

take into account possible effects on surrounding areas.

Satellite imagery used

This study relied on freely available satellite imagery through the Sentinel-series of the European

Copernicus programme, namely Sentinel-1 and Sentinel-2, and thus combined radar and optical data.

The fusion of both data sources has been shown to improve biomass estimates, especially in open

woodlands as in Northern Uganda (Joshi et al. 2016). The use of very high resolution commercial

optical imagery had been considered, but there was no adequate imagery available covering the full

settlement area neither for the end of 2016 nor the beginning of 2017.

9

Radar data

Sentinel-1 is a C-Band SAR satellite. Sentinel-1 data was selected for the period of 1/1/2016 until

15/3/2017, considering both single and dual polarised data. A total of 16 images for Track 102 was

found and has been subsequently downloaded as Level-1 GRD product from Alaska Satellite Facility

server using the Open Foris SAR toolkit interface on the SEPAL platform (https://sepal.io). This data

has been used to create a time-series stack and the subsequent application of a multi-temporal

speckle filter.

Afterwards two separated timescan products were elaborated at 10 m spatial resolution. Since VH

polarised imagery was not available for all the dates of interest, only the VV polarised bands were

considered in this step. The procedure itself consists of taking the time-series stack and creating

descriptive statistics such as the minimum, maximum and average value as well as the standard

deviation and coefficient of variation, calculated in the temporal domain. Dry season images were

chosen for the two timescan products (four images for the beginning of each of 2016 and 2017)

because the SAR signal is not just sensitive to structural attributes, but also to moisture effects. Since

woody vegetation is characterized by low temporal variation and relatively high backscatter values, a

correlation between the timescan products and the biomass can be expected.

Optical data

Sentinel-2 is an optical sensor which captures spectral reflectance of the wavelengths characteristic

for the differentiation of vegetation. Image mosaics (greenest wettest pixel closest to the target date)

at a spatial resolution of 10 m were acquired through Google Earth Engine API for February 2016 and

January 2017 – before and after the establishment of the settlement. February 2016 was chosen

because of seasonal similarities to January 2017.

Figure 5: Sentinel-2 and Sentinel-1 data for the area of interest for beginning of 2017

10

Figure 5 (Left): False colour composite of Sentinel-2 mosaic for January 2017, including boundaries of

settlement area and 5 km buffer around. Band 8 (near infrared) is displayed as red, band 4 (red light)

as green and band 3 (green light) as blue. Green vegetation shows as red in this image, grasslands (dry

at this time of year) show as light green/brown, and dark areas are burnt areas.

Figure 5 (Right): Pseudocolour image of Sentinel-1 Gamma0 VV minimum backscatter intensity

timescan layer for January/February 2017. Dark green areas indicate high woody biomass, white areas

no woody biomass.

Image processing for LULC classification and direct biomass estimates

Both Sentinel-1 and Sentinel-2 data were processed to the same spatial extent and resolution (~10 m)

and clipped to a 5 km buffer around the settlement boundary.

The following processing steps were undertaken for each time step:

Processing of Sentinel-1 imagery to same spatial extent and resolution (~10 m) of Sentinel-2 imagery;

For 2017, use of bands 2, 3, 4, and 8 for Sentinel-2 imagery and all four available bands from Sentinel-1 multi-temporal data (mean, min, max, standard deviation). For 2016, use of only bands 2, 3, 4, and 8 for Sentinel-2 because Sentinel-1 data contained artefacts that presented biased results;

Unsupervised classification (50 clusters) run over the imagery for each time period (oft-k-means);

Sieved the clusters to a minimum size of 8 pixels (gdal_sieve.py);

Mean spectral signature calculated for each unique cluster (oft-stat);

RA supervised classification (R package random Forest) run for LULC classification;

Computed statistics of resulting classification images for both settlement area only and buffer area of 5 km around the settlement boundary.

Several training data sets were tested for the supervised LULC classification:

Points visited during the field mission only;

Combination of points from field mission and manually added points;

Hand-drawn polygons.

Dataset 1 was insufficient to derive a sound classification. Dataset 2 provided very good classification

results for January 2017, whereas dataset 3 was deemed most appropriate for February 2016.

Change detection

Independent change detection was performed using the IMAD algorithm (oft-imad.py) between the

two time periods on the Sentinel-2 imagery. This ensured that zones of changes were properly

detected, taking into account issues related to seasonality and the presence of clouds and shadows.

11

Tools and software used

Field data was collected using Open Foris Collect Mobile, with its survey developed and database

handled in Open Foris Collect.

Remote Sensing data was processed on SEPAL (System for earth observations, data access, processing

& analysis for land monitoring, https://sepal.io). SEPAL is a cloud-based system developed by FAO that

allows the user easy access to satellite imagery and powerful processing capacities. Details about

software and tools used are listed in Table 1.

Description Software/tool used

Field data collection Open Foris Collect Mobile

Field data database Open Foris Collect

Sentinel-2 acquisition Google Earth Engine API

Sentinel-1 acquisition Open Foris SAR toolkit

Working environment SEPAL Ubuntu 16.04

Processing code Bash 4.3.46

Spatial data processing Open Foris Geospatial Toolkit 1.26.6

Calculations R 3.3.2

Visualization and spatial editing (on local machine) QGIS 2.18.3

Table 1: Software and tools used

1.3.5 Definitions

LULC classes

Wood resources are distributed unevenly between land use/land cover (LULC) classes in the Area of

Interest (AOI). The LULC classes used in this study are similar to the ones used for the National Biomass

Study (NBS) of Uganda (Forest Department 2002). The main classes found in the field inventory were

closed and open woodlands, bushland, grassland and cultivated land. For the LULC mapping, also

bare/open land, water and burnt areas were added. The field plots were unevenly distributed among

the LULC classes with number of sample plots per class ranging from 3 to 20 (Table 5 in the results

section 3.4).

These classes are described as follows according to the NBS (Forest Department 2002, pp. 17-19):

Woodlands – “Wooded areas where trees and shrubs are predominant. There are wet and dry types.

The wet type occurs as a zone along wetlands (riverine forest) and the dry type is found on grass-

covered upland areas. To qualify as woodland the average height of the trees must exceed 4 m”.

Woodlands can be distinguished into closed and open woodlands by their canopy cover. Closed

woodlands have a canopy cover of > 40% whereas open woodlands have one of 10 to 40%.

12

Bushlands – “Refers to vegetation dominated by bush, scrub and thicket growing together as an entity,

but not exceeding an average height of 4 m”.

Grasslands – “Rangelands, grazing grounds, improved pastures and natural savannah grassland.

Various trees - bush/woody vegetation frequently occur on this land, but grass dominates the

landscape”.

Cultivated land/farmland area – “Scattered trees are frequently found in the vicinity of the

homesteads. Examples include fruit trees and various multipurpose trees integrated in the farming

system (agroforestry). Farmland areas including small holder subsistence farm units cover 50-90% of

the land cover of Uganda. The cropping systems include mono- and mixed cropping”.

Bare/open land includes bare soil without or very limited vegetation cover and built-up areas and

settlements without significant amounts of vegetation.

Wood resources

The focus of the study was on classes with potential woodfuel resources. Both bushland and grassland

were found to have very low aboveground biomass stocks and few sample plots and were therefore

aggregated into one class.

The field data captured three main sources of wood resources:

- aboveground biomass (twigs, branches and stems) of shrubs and trees;

- deadwood (both in standing trees and on the ground);

- saplings as a proxy for regeneration potential.

13

2. Woodfuel demand results

The woodfuel demand in Bidi Bidi settlement is concentrated at household level. There are schools in

the settlement but feeding programs are not yet in place. Some fuelwood is used for cooking also at

the refugee reception centre though at very low levels compared to the total consumption at

household level. Furthermore, wood is also used as building material, thus an estimate of woody

biomass extraction for housing the refugees was also considered.

2.1 Population and household information

The current population of Bidi Bidi settlement is estimated at 272 206 refugees (UNHCR - Level 1

Registration to 19 December 2016). All Bidi Bidi settlement dwellers come mainly from five different

ethnicities (i.e. Madi, Peri, Lotuko, Acholi, Dinka) of South Sudan; some speaking similar languages of

Northern Uganda like the Acholi and Madi.

A total of 72 households were randomly selected with 68 females and 4 males interviewed; 11 HH

were male headed; 61 HH were female headed. The majority of people interviewed were female

between 25 and 40 years of age (68%); 38 (53%) were heads of household while the others were

spouses (wife) 25%, daughter (2.8%) and 19% were not specified.

Size of households

The estimated size of household in this study was based on the household survey on a sample of 72

households. The majority of households interviewed range between four and seven members (Figure

6) and the average number of members in a household is estimated at six.

Figure 6: Average household size

Livelihood

More than 66% of the households in the refugee settlement engaged in crop agriculture before the

crisis; only two households had two different sources of income; 22% had been engaged in some

0369

1215182124

1 2 3 4 5 6 7 8 9 ormore

Household size of the sample

%

People in the household

14

business related to pastoralism, fishing, charcoal production, cooking food, masonry, poultry keeping

and e education; 5% were agro-pastoral, while 3% had members who worked as health workers.

It is important to note that although all 72 households did not mention any number of income/wage

earners, a few households stated having some occasional sources of income. Just over 26% reported

selling food and one household was involved in the sale of fuelwood. The majority of households do

not have any source of income.

Focus group discussions also confirmed that people in the refugee settlement generally do not have

sources of income; the few who try are engaged in petty trade of selling food items and a few

necessities. Makeshift markets were also observed in Bidi Bidi settlement. A few other refugees had

been engaged in a one-time contract of making fuel-efficient devices by NGOs (FGD- F: 6; Zone 4). A

few refugees own some livestock which they sell for a small amount of money.

Some of the refugees have resorted to negative coping mechanisms to meet other household needs,

for example selling solar lamps at UGX 20 000 – 30 000 per lantern (approximately US$ 5-8), and selling

goats and sheep at about 40–50% of market price.

The income opportunities available for refugees also include bread/mandazi (a type of local fried

bread) making, cooking food, for instance a maize and beans-based meal, or petty trade of soap, salt

and sugar, but developing these opportunities need some cash as capital investment which the

refugees do not have under current circumstances.

The majority of households in the host communities depend on natural resources for their main

incomes: agricultural production, selling fuelwood and charcoal and grass for brooms. However, they

have perceived a reduction in the amount of trees in the recent past.

Currently, there is competition for fuelwood and grass for grazing animals and thatching houses

between the refugees and host communities. This can potentially cause conflict and tension.

2.2 Woodfuel consumption for cooking and heating

In this assessment refugee households were asked about the type of fuel used for cooking/heating

and to estimate the average amount of woodfuel they used for cooking and heating in a day. According

to information gathered from the household survey, a considerable amount of fuelwood is consumed

for the preparation of beans that require a long cooking time. Fuelwood is also used to heat water for

children’s baths and to provide heat in the cold mornings or evening. The fuelwood was then weighed

and recorded for both the daily fuelwood consumption for cooking/heating and the specific task of

cooking beans.

The majority of refugee households use only fuelwood (97%) for cooking/heating and only 3% use

charcoal for cooking (Table 2). Currently the majority of the households (96%) collects fuelwood

directly from the surrounding forest and other woodlands; only 4% purchase from the market and 2%

get fuelwood from both the forest and the market.

15

Table 2: Fuel type used in the settlement

Households in the surrounding host community also mostly use fuelwood for cooking, heating water

as well as for brick making to be used as building materials – a practice that is not permitted within

the refugee settlement area. Inside Bidi Bidi settlement, a household receives maize and red sorghum

grain, beans and some other basic food items every month. Usually each household cooks and lights

a fire twice a day.

The average fuelwood consumption in a day for cooking/heating needs, other than cooking beans,

corresponds to 12 kg per household that is equal to 2 kg of fuelwood per day per person. It has been

recorded through household interviews that beans take a very long cooking and yet 83% of refugee

households cook beans at least 4 times in a week. If each household cooks the beans an average of 4

times in a week and distributing the fuelwood consumption on a daily basis, the estimated fuelwood

consumption per person per day for cooking beans is an average of 1.5 kg.

Thus, the overall average fuelwood consumption for cooking/heating is estimated at 20.5 kg per day

per household that corresponds to 3.5 kg of fuelwood per day per person. This figure is considerably

high if compared to similar situations. For instance in literature the fuelwood consumption for cooking

in displacement settings ranges from between 0.7 kg to 3 kg per person per day (Gunning, 2014). The

long time required for cooking the beans is identified as a driver of such high fuelwood consumption.

In addition, it was observed that refugee households occasionally use wet/green wood, which entails

a lower efficiency rate in burning fuelwood. A more in-depth analysis of the current technologies and

practices for cooking is needed to better understand the main drivers of fuelwood consumption.

By scaling up the results from the sample household survey, for 272 206 refugees registered in Bidi

Bidi settlement, the rate of fuelwood consumption is estimated approximately at 952 tonnes per day

with a demand of over 347 480 tonnes of fuelwood per year. Since the displacement setting was

opened (over 7 months ago), the total fuelwood extraction is estimated at about 199 920 tonnes.

Fuel type # HH %

Fuelwood 70 97.2

Charcoal & fuelwood 2 2.8

Total 72 100

16

Figure 7: Members of the refugee settlement estimating and weighing the daily wood consumption

Charcoal

Refugees were not permitted to produce charcoal by the local authorities although they are free to

purchase and use it. During the data collection, the impression was that many were not free to disclose

that they use charcoal out of fear that their use of charcoal would be considered illegal. The household

survey shows that very few of them (3%) use charcoal for cooking with an average consumption of 2.7

kg per day that corresponds to 0.4 kg of charcoal per person/day. Therefore, taking into account 3%

of total refugee population using charcoal, the total charcoal consumption in Bidi Bidi is estimated to

3.2 tonnes per day and 1 170 tonnes per year. Assuming a conversion efficiency of fuelwood to

charcoal of 20%, the total fuelwood consumption is estimated to be approximately at 5 840 tonnes

per year. Since Bidi Bidi settlement was opened the charcoal consumption is estimated to be 672

tonnes, which corresponds to 3 360 tonnes of fuelwood.

Refugees are experiencing a lack of fuelwood and have started devising ways of coping with the

scarcity:

“... sometimes exchange food for charcoal; before crisis most refugees were using fuelwood

but now even smaller pieces of wood, twigs and maize cobs are used”. (Focus Group Discussion,

F&M Refugees)

Other uses of woody biomass

It was noted that wood has been cut down by the refugees also for building shelters and other

household constructions. According to the field measurements carried out in the field, approx. 0.9 m3

of wood poles has been used by each household for house construction. Therefore, the estimated

woody biomass used for this purpose by the 101 500 households registered in Bidi Bidi settlement is

91 350 m3 that corresponds to 63 945 tonnes (assuming an average wood density of 700 kg/m3).

17

Results of woodfuel demand and other woody biomass uses

Since the establishment of Bidi Bidi settlement, the estimated total woody biomass used for fuelwood,

charcoal and building purpose is 267 225 tonnes. Taking into consideration the energy needs of

refugees for cooking and heating in Bidi Bidi settlement, the estimated fuelwood demand is 353 320

tonnes per year (including charcoal on a fuelwood basis) (Table 3).

Baseline wood consumption

Total daily consumption (t)

Total annual consumption (t)

Total consumption since the establishment of

Bidi Bidi (7 months) (t)

Fuelwood 952 347 480 199 920

Charcoal* 16 5 840 3 360

Building material 63 945

Total 968 353 320 267 225

* Expressed on fuelwood basis assuming a conversion efficiency of fuelwood to charcoal of 20%

Table 3: Baseline wood consumption of Bidi Bidi settlement

2.3 Technologies and practices for cooking

The assessment team asked respondents about their experiences with cooking/heating devices, type

and problems experienced before fleeing to Uganda and their current situation. The team also

observed and took pictures of some of the cooking devices used in the settlement and host community.

There are more households in Bidi Bidi settlement using the traditional 3-stone fire (56%) compared

to before the crisis (32%). However, a considerable proportion of households uses an improved mud

stove for fuelwood (40%). From observations however, many households who have the mud stove

still use the 3-stone fire. Additionally, 3% of households use the mud stove for charcoal and only 1.4%

households is currently using the metal charcoal stove compared to 17% before the crisis (Figure 8).

Figure 8: Type of cookstoves used prior to crisis and at the present

0%

10%

20%

30%

40%

50%

60%

3-stone Mud stove forfuelwood

Mud stove forcharcoal

Metal stove forcharcoal

Household cookstoves

Prior to crisis Current situation

18

It was observed among the households that the mud stoves for fuelwood used in Bidi Bidi settlement

do not have a standardised make. This means the stove energy efficiency might be different according

to the design used. The household survey results do not actually show a marked difference in

fuelwood consumption between households of the same size using the 3-stone fire and the mud stove.

However, during the focus group discussions, respondents indicated that there is a difference in the

fuel consumed, with the mud stove requiring less. Further investigations are recommended to verify

the energy efficiency of these mud stoves.

Many households in the host community also use the traditional 3-stone fire and a few use some fuel-

efficient devices. However, this assessment did not investigate the different kinds of devices in the

host community.

Figure 9: A community worker cooking beans in her household using the traditional 3-stone fire

From the direct interviews conducted with household in the settlement, it was revealed that many

refugees have knowledge and skills to make and use fuel-efficient mud stoves but with different

adaptations in the design. Some people have received training on the construction and value of fuel-

efficient stoves but a good number of people did not receive training but have copied the mud stove

production from a neighbour or relative.

“My 16 year old daughter constructed the fuel-efficient stove after seeing it at the neighbours”. (FGD-

mixed Refugees, F:3; M:2 Block 11)

“I saw already one mud stove and I thought I could make it myself and that’s how I have and use one; I

first made the bricks and then prepared mud after bricks had dried”. (FGD- F: 6; Zone 4)

“Only three households in the village are using the fuel-efficient stove, they now have time attend to

other household chores and making craft for income”. (Focus Group Discussion, Host Community)

19

The respondents stated that the mud stove protects the fire from the wind effect, requires less wood

than the 3-stone fire and consequently cuts down the fuelwood collection trips saving time for other

productive work.

Figure 10: An improved mud stove used by a refugee household

Figure 11: A woman cooking posho (maize meal) on an improved mud stove (Two-Pot Rocket Lorena)

Problems experienced by the households with different cooking methods

In the current situation the most common problems experienced by refugee households using the 3-

stone fire is the high fuelwood consumption and smoke inhalation. Few households have experienced

high fuelwood cost and undercooked food in the settlement.

20

The host community also experienced the problems of smoke inhalation, high fuelwood

consumption but not undercooked food except when they purchased beans from refugees. This fact

highlights that perhaps the beans provided to refugees are old and therefore require more cooking

time in comparison to the beans produced locally.

“... sometimes food is undercooked if beans are purchased from refugees; lots of fuelwood used” (Luzira

Village Group Discussion, Host Community)

Generally 42% of the households interviewed use a type of mud stove for cooking and they mostly

mentioned the same problems of high fuelwood consumption and smoke inhalation; the fixed mud

stove is not moveable during the rainy season and it may break within one week because of the rain.

The host community who have been trained through GIZ and UNHCR programmes on the construction

of portable mud stoves pointed out some of the disadvantages: it needs to be protected from rain,

but it is heavy to move; it cuts off heat and you are not able to warm yourself in the cold; it can be

stolen; it needs skills to construct; and requires constant maintenance (Focus Group Discussion, Host

Community).

Figure 12: Refugees and host community members working on a CDO project to produce energy saving devices

Compared to the refugee community the adoption of improved fuel-efficient devices is still very low

in the host community; out of the village group discussion with over 30 women and men only three

households were using a fuel-efficient stove for cooking.

From field observations, many households in the Bidi Bidi settlement have the skills to make a fuel-

efficient stove from locally available materials and some of the household members have been

engaged in training and a one-time off paying contract of construction of the portable improved mud

stoves by supportive programme from UNHCR, NGOs and local authorities.

21

A number of people have received training on fuelwood saving techniques both in the refugee and

host community. Some refugee women have received a one-day training, while members of the host

community had just received a one-week training to community groups.

The opportunity of having already trained people can help in sharing knowledge in construction and

use of improved stoves and promoting awareness to get fuel-efficient stoves in every household.

Food preparation methods

In Bidi Bidi settlement the majority of households do not soak or process the beans (74%) before

cooking in any way; 25% stated they soak beans, however key informants explain that “soaking” is

typically for 15 minutes, which is followed by grinding (breaking the beans in smaller pieces) before

cooking.

Table 4: Cooking practices registered in the settlement

According to information gathered through the household survey, beans are a main food in all refugee

households, with the longest cooking time and the highest fuel consumption for their preparation in

comparison to any other type of food. It has been recorded through the household interviews that

beans require a very long cooking time and yet 83% of refugee households cook beans at least 4-5

times in a week.

Soaking overnight greatly reduces the cooking time and fuelwood consumption. Soaking pulses such

as beans also has the additional benefit of increasing their nutritional benefits, and allowing nutrients

to be better absorbed by the body.1 This would also have an impact on fuelwood consumption.

However, many people do not like soaking beans stating that they were not palatable or smell bad;

they would rather go the length of cooking for longer hours than soaking beans.

1 http://www.fao.org/3/a-i5388e.pdf

Preparation of beans #HHs %

# HH who “soak” beans 18 25.0

# HH who don’t soak beans 53 73.6

Missing value 1 1.4

Total 72 100

22

2.4 Challenges and risks to access woodfuel

A number of social and economic vulnerability issues were pointed out as challenges associated with

the collection and use of fuelwood: carrying heavy loads, walking long distances, assault/rape, lack of

tools, fuel costs for some and the burden of household chores.

Fuel collection distance - The responsibility of fuelwood collection falls entirely on women and girls

above 15; they go either with others or on their own. This involves walking up to wherever the

fuelwood may be, cutting it and transporting it back home. Most women walk 2-3 km to collect

fuelwood but some walk up to 5 km and some go up to host community areas to look for fuelwood.

They also expressed that the distance has steadily increased in the past months as the availability of

wood close to the settlement has diminished. Moreover, given to the low availability of woody

biomass, people are increasing the use of green or wet wood and the time take in cutting as well as

drying fuel is a barrier to more efficient cooking practices.

“Women collect fuelwood almost every day; taking 2-3 hours; it is tasking and tiring”. (FGD-

mixed Refugees, F:3; M:2 Block 11)

“...when we came to Bidi Bidi there were many trees around the settlement, now we have to

walk longer distances for round trip it takes 2-3 hours”. (FGD- F: 6; Zone 4)

Wild animal threats – A number of women talked of facing fears of wild animal attacks during

fuelwood collection.

Assault/rape – Refugee women face a number of challenges in looking for fuelwood; they face fears

of being assaulted by the host communities and this may be in form of physical beating, verbal abuse,

or even rape:

“Yesterday two girls were chased by host community members complaining that they should

not collect fuelwood from here”. (FGD-mixed Refugees, F:3; M:2)

“Sometimes people in the host community prevent refugees from collecting fuelwood and

producing charcoal”. (FGD- F:6)

Incidents of rape are reported not only as performed by members of host communities but also within

the refugee settlement.

Lack of tools - According to a key informant and FGDs refugees lack tools such as axes to collect

fuelwood.

“Lack of axes, the machetes are not able to cut, they have broken handles, not easy to use or

mend (lack materials to mend the handles)” (FGD-mixed Refugees, F:3, M:2; Block 11)

However, this might be an indication that the number of small trees has decreased and individuals are

targeting bigger trees, which require more powerful tools.

Poor dietary practices – Before the crisis, refugees cooked a variety of foods including: cassava, millet,

sorghum, maize, sweet potatoes, yams, pumpkin, greens, cabbage, meat and cowpeas. Currently the

only foods mentioned in the refugee settlement were posho/maize and beans with a few having daga

23

fish. Nutrition/dietary needs are compromised within many households only eating posho and beans,

having only two or less meals a day for most (51%) with just 47% of households having three meals in

a day. A typical refugee household in Bidi Bidi will have porridge for breakfast, posho/maize and beans

for lunch. Many households have cut down the meals to two in a day because of the uncertainty of

the food rations to come. Sometimes refugees cope by exchanging food for fuelwood or cooking once

in a day i.e. cook food enough for lunch and keeping the same food for the evening meal.

“Currently most households cook beans, posho, porridge which has been suspended because

the rations have not yet been received”. (FGD-mixed Refugees, F:3; M:2 Block 11)

Conflict and tension - During focus group discussions, a number of issues came up that indicate some

tension in the refugee or host community: competition on the same natural resources, unfulfilled

agreements, shared services and some miscommunication. As natural resources start diminishing,

tensions are growing between the host and refugee communities, particularly since the host

community can no longer access resources within Bidi Bidi settlement. Competition for collecting

fuelwood between the host community and refugees can cause conflict and tensions if not addressed.

The designation of Bidi Bidi as a refugee settlement has led to competition for local natural resources

such as wood for energy and building material as well as grass for thatching and grazing; it has also

cost the host community a part of their grazing lands and sources of thatching material. Both the host

and refugee communities have complained that the other burns the grass or takes their cut fuelwood.

A few misunderstandings have also sprung up causing some tension between the refugee and host

communities on the use of land and water. The refugee community, for example, believes that the

land on which they were settled was purchased and paid for them while the host community lent the

land in exchange for services agreed with the local authorities.

Land and tree tenure – for the establishment of woodlots or tree plantatations there may be

challenges associated to the land tenure system in Yumbe. The majority of land nationwide is under

customary tenure; although farmers on customary land are protected by the law and enjoy security

of tenure there is a need to discuss issues pertaining to tenure of land and trees that may be planted

on different land tenure systems within Yumbe District. Issues need to be agreed upon, with a gender

perspective and represented in a formalized document (refer to VGGT note in recommendations).

The practice of involving end users of both communities (hosts and refugees) in the selection of

multipurpose trees to be planted – already encouraged by UNHCR in reforestation programmes2 –

could alleviate tension, particularly if there are gains beyond fuelwood to be made (e.g. food, fodder,

and shelter) which can meet both subsistence and income needs.

2 http://www.unhcr.org/protect/PROTECTION/438724c42.pdf

24

3. Woodfuel supply results

3.1 Results from the field measurements

Only aboveground biomass and deadwood was targeted in this study as primarily source of woodfuel

and it is the only resource that can be directly inferred from available satellite imagery.

The potential woody biomass for fuel of each LULC class differs significantly in terms of aboveground

biomass, deadwood and number of saplings (Table 5 and Error! Reference source not found.). Similar

verage aboveground biomass in cultivated land and open woodland is in line with findings from NBS

and other field inventory data from Uganda.

LULC class Number of plots surveyed

Trees per ha Aboveground biomass (t/ha)

Deadwood (t/ha)

Saplings per ha

Closed woodland 3 1209 ± 2101 56.96 ± 44.75 1.62 ± 5.46 265 ± 900

Open woodland 20 647 ± 1310 9.23 ± 13.57 17.58 ± 144.73 368 ± 1090

Bushland and grassland

6 394 ± 1095 2.02 ± 5.37 0.64 ± 1.56 133 ± 472

Cultivated land 3 367 ± 1112 12.49 ± 7.38 0.56 ± 1.89 66 ± 225

Table 5: Woodfuel potential of each LULC class including 95% confidence intervals

A total of 43 tree species was recorded in the field data (Annex 1 for complete list of tree and shrub

species). The dominant tree species are Acacia hockii (102 trees), Combretum collinum (67 trees),

Grewia mollis (42 trees) and Combretum fragans (28 trees). A total of 4 shrub species were recorded;

for some shrubs the species could not be identified, mainly due to burning (Annex 1).

Figure 13: Aboveground biomass distribution between LULC classes. The boxplot depicts the median, first

and third quantile and largest and lowest values

25

Estimating growth rates in terms of mean annual increment requires repeated measurements in time

which were not available for Bidi Bidi settlement. However, the NBS provides such figures as national

average and for agroecological zones (Yumbe District falls into the semi-moist lowland zone) (Forest

Department 2002). The growth rates differ widely between LULC classes (Table 6).

LULC class

National mean annual increment

(t/ha)

Semi-moist lowland mean annual increment

(t/ha)

Woodland 5.0 4.2

Bushland 1.0 0.3

Grassland 1.0 1.2

Subsistence Farmland

1.0 1.7

Table 6: Annual increment in t/ha (air-dry matter) from National Biomass Study 2002, as national averages and for moist lowlands (Forest Department 2002)

3.2 Mapping the distribution of LULC

In addition to the four LULC classes present in the field data, three other classes were mapped:

open/bare land, water and burnt areas. Mapping burnt areas separately was necessary because the

optical Sentinel-2 imagery was from the dry season where burnt areas regularly occur in this part of

the country. Open/bare land and water are not present in the field data because they have no

aboveground woody biomass.

The spatial distribution of LULC classes inside the settlement and 5 km around the boundary for 2017

is depicted in Figure 14. Calculated areas per LULC class are available in Table 7.

26

Figure 14: LULC 2017 for Bidi Bidi settlement and 5 km buffer around the settlement boundary

LULC class Area within Bidi Bidi settlement boundary (ha)

Total area including 5 km buffer (ha)

Closed woodland 3 356 8 935

Open woodland 39 492 69 164

Bushland and grassland 18 166 34 185

Cultivated land 10 598 29 829

Bare/open land 3 249 5 101

Burnt areas 4 905 14 352

Water 1 126

Total 79 765 161 693

Table 7: Area of each LULC class within settlement area and in total area with 5 km buffer around the settlement boundary

27

3.3 Mapping changes in LULC

Changes for such a short time frame as in this study (11 months from February 2016 to January 2017)

are very difficult to detect from satellite imagery because they get overridden by changes from natural

(seasonality) and anthropic (agricultural burning practices) origins that are of larger order of

magnitudes (in terms of area) and spatially spread within a complex landscape. With the IMAD change

detection, only changes inside the zones, mainly related to the construction of roads, and seasonal

differences of burnt areas could be detected.

For this purpose, all areas within the settlement boundary which showed bare/open land in 2017 and

other LULC in 2016 and which were above 5 pixels (~ 500 m2) in size were assumed to be changes due

to the camp establishment. This resulted in an estimated total area of 2 295 ha converted from other

LULC to bare/open land through the establishment of the settlement. This area was mainly comprised

of open woodlands with 1 166 ha (Table 8).

LULC class 2016 Area (ha)

Closed woodland 8

Open woodland 1 166

Bushland and grassland 680

Cultivated land 119

Burnt areas 322

Total 2 295

Table 8: Area of LULC in 2016 converted to bare/open land in 2017 through camp establishment

3.4 Biomass mapping

To estimate and map the available wood resources for woodfuel, the mean aboveground biomass is

applied to the respective LULC class. As burnt areas were not measured as such in the biophysical field

inventory, they were assumed to have the same aboveground biomass stock as bush- and grassland

because these LULC types usually burn.

Table 9 shows for each LULC class the area (ha) both within the settlement boundary and including

the 5 km buffer, and the mean aboveground biomass (t/ha). These two are multiplied to derive total

available stock of aboveground biomass. The distribution of the biomass stock is shown on the map in

Figure 15.

28

Figure 15: Distribution of wood resources in Bidi Bidi settlement (Including a 5 km buffer) in January 2017

The total aboveground biomass stock is estimated at 734 614 t for the settlement area and at

1 617 953 t including the 5 km buffer area around the settlement boundary. Most aboveground

biomass can be found in open woodlands, but also closed woodlands contribute significant amounts

despite their limited coverage in terms of area.

29

LULC class Area within settlement boundary (ha)

Total area including 5km buffer (ha)

Mean AGB (t/ha)

AGB stock within settlement (t)

AGB stock within settlement plus 5 km buffer (t)

Closed woodland

3 356 8 935 56.96 191 130 508 965

Open woodland

39 492 69 164 9.23 364 513 638 380

Bushland and grassland

18 166 34 185 2.02 36 695 69 054

Cultivated land

10 598 29 829 12.49 132 369 372 562

Bare/ open land

3 249 5 101 0 0 0

Burnt areas

4 905 14 352 2.02 9 908 28 992

Water 1 126 0 0 0

Total 79 765 161 693 NA 734 614 1 617 953

Table 9: Area, mean AGB and total biomass stock estimates for each LULC class

The establishment of the settlement had a detectable impact on 2 295 ha, which have a total

aboveground biomass stock of 14 725 t (Table 10).

Table 10: Area of LULC in 2016 converted to bare/open land in 2017 through camp establishment and total

aboveground biomass associated with these areas

LULC class 2016 Area (ha) Total aboveground biomass lost (t)

Closed woodland 8 450

Open woodland 1 166 10 763

Bushland and grassland

680 1 373

Cultivated land 119 1 488

Burnt areas 322 651

Total 2 295 14 725

30

4. Possible scenarios and recommendations

4.1 Possible scenarios

Several possible interventions have the potential to improve energy access, minimize environmental

impacts in Bidi Bidi, and bring additional benefits such as livelihood opportunities from multipurpose

tree planting and time spared from reduced collection of fuelwood. Based on the integration of the

results from woodfuel demand and supply assessment, the following future scenarios are developed:

1) Scenario Business as Usual (BAU); 2) Scenario with improved cookstoves; and 3) Scenario with tree

plantations and improved stoves.

(1) Scenario Business as Usual. The rate of fuelwood consumption is estimated at 952 tonnes per day

or 347 480 tonnes per year and the total aboveground biomass stock in the settlement area is

734 614 tonnes with an annual aboveground biomass growth of 33 300 tonnes per year in the

settlement area. By considering only the settlement area, the annual deficit of woody biomass

growth in the targeted area is 314 180 tonnes. The estimated total aboveground biomass stock

and growth from trees and shrubs within the settlement area, in a hypothetical worst case scenario,

can supply fuelwood to the current refugees in Bidi Bidi for 3 years (Table 11Table 10). This would

however be at the cost of a full depletion of the Bidi Bidi settlement area in terms of aboveground

biomass, representing an issue that needs to be addressed as soon as possible as refugees already

compete with the local communities for the scarce resources outside the settlement boundary area.

Year Total fuelwood consumption

(t yr-1)

Total AGB stock (t)

Annual AGB growth (t yr-1)

Annual AGB loss (t yr-1)

1 347 480 734 614 33 300 -314 180

2 347 480 420 434 19 058 -328 422

3 347 480 92 012 4 171 -343 309

4 347 480 --- --- ---

Table 11: Woodfuel supply and demand in BAU Scenario

(2) Scenario with improved cookstoves. Under this scenario woodfuel demand is reduced by using

fuel-efficient cookstoves. Taking into consideration the baseline situation, 56% of households are

currently using the 3-stone fire against to 42% using improved mud stoves, this scenario assumes

that 100% of households will shift to an improved mud stove with 20% thermal efficiency.

Therefore, assuming efficiency of 3-stone fire at 10%, the utilization of improved stove instead of

3-stone fire by 56% of households can result with about 30% of fuelwood savings compared to the

BAU Scenario. The total fuelwood demand in Bidi Bidi settlement is then estimated about 243 000

tonnes per year. Under this scenario with the introduction of improved cookstoves, women will

also spend less time collecting fuelwood with more time available for other activities. As the

estimated annual aboveground biomass growth within the settlement area is about 33 300 tonnes

per year, the annual deficit of woody biomass growth under this scenario would be 209 700 tonnes

(year 1). Therefore, according to the estimated total aboveground biomass stock and annual

31

growth from trees and shrubs within the settlement area, the total aboveground biomass

available can supply the current refugees in Bidi Bidi for 4 years in this scenario (Table 12). This

would however be at the cost of a full depletion of the area in terms of aboveground biomass,

representing an issue that needs to be addressed as soon as possible. Therefore, the utilization of

improved mud stoves by the all refugee households can only delay the deforestation and land

degradation rate resulting from the woodfuel demand for cooking.

Year Total fuelwood consumption

(t yr-1)

Total AGB stock (t)

Annual AGB growth (t yr-1)

Annual AGB loss (t yr-1)

1 243 000 734 614 33 300 -209 700

2 243 000 524 914 23 794 -219 206

3 243 000 305 708 13 858 -229 142

4 243 000 76 566 3 471 -239 529

5 243 000 --- --- ---

Table 12: Woodfuel supply and demand in Scenario with improved cookstoves

(3) Scenario with tree plantations and improved stoves. Another scenario for reducing the

environmental impacts resulting from the woodfuel demand is to increase the woody biomass

production by planting trees in combination with the introduction of improved cookstoves

(Scenario 2). When considering common productive woodlots in Uganda with a MAI that ranges

between 20 and 26 m3 per hectare, and assuming an average wood density of 600 kg/m3, a

biomass expansion factor of 1.5 (to include also bark and branches to the MAI), the total

aboveground biomass increment would range between 18 and 23.4 tonnes per hectare per year.

To compensate the annual biomass loss of 219 206 tonnes (after 2 years) and assuming the tree

plantation becomes productive in the year 3, between 9 368 and 12 178 hectares would need to

be planted with fast growing species that corresponds to 12-15% of the total land area of Bidi Bidi

settlement (798 km2 within the settlement boundary) (Table 13). Each household would need a

minimum area of 50 m x 50 m, dedicated exclusively to growing timber for fuelwood. If this number

of hectares could be planted, this would provide a sustainable source of fuelwood for the existing

population, however this involves considerable costs, land ownership issues, expertise, forest

management, etc. Forest management should focus on maximizing the biomass production in the

short time and tree density can be drastically increased to reach the optimum growth per unit of

surface. Short rotation coppice management can regularly produce fuelwood for long time periods.

Year

Total fuelwood

consumption (t yr-1)

Total AGB from existing

resources (t)

Annual AGB growth from

existing resources (t yr-1)

Min. tree plantation

area (ha)

Annual AGB growth from

tree plantation (t yr-1)

Annual AGB loss

(t yr-1)

1 243 000 734 614 33 300 9 368 0 209 700

2 243 000 524 914 23 794 9 368 0 219 206

3 243 000 305 708 13 858 9 368 219 206 0

4 243 000 286 277 12 977 9 368 219 206 0

5 243 000 265 965 12 056 9 368 219 206 0

Table 13: Woodfuel supply and demand in Scenario with tree plantations and improved stoves

32

It is important to note that experience in reforestation programmes suggests that in order for such

an option to be sustainable and attractive to the end user, tree species should be selected

according to the needs of these users – and multipurpose trees that might fulfil additional needs

(e.g. food, animal fodder or shade) should be integrated in this option3.

The likelihood is that a combination of these interventions will be needed, effective immediately,

to reduce the high woodfuel demand from the refugees of Bidi Bidi settlement and to support a

sustainable woodfuel supply. These include the adoption of fuel-efficient stoves and energy

savings cooking practices at the household level, establishing multipurpose tree planting

programmes, developing forest management plans and promoting woodfuel alternatives (e.g.

agricultural residues, LPG). Such interventions would have the added value of contributing to

climate change mitigation through carbon sequestration and reduction of CO2 emissions. The

carbon sequestrated by forest plantations and the emission reductions from improved cookstoves

could also provide benefits through carbon credits.

4.2 Recommendations

Fuel-efficient cookstoves and energy saving measures

Fuel-efficient cookstoves should be promoted to reduce fuelwood demand with about 30% of

fuelwood savings compared to the BAU Scenario. It is important to increase effectiveness of

programmes to improve the use of fuel-efficient stoves and the adoption of energy efficient cooking

practices through engaging relevant expertise to design community awareness and engagement

programs. A participatory approach should be promoted for planning and monitoring these activities

for the refugee and host communities.

It is important to keep in mind that the selection of appropriate fuel-efficient cookstoves has to take

into consideration several specific factors such as safety, availability of local fuel resources, funding

levels, cooking habits, user acceptability4. An in-depth assessment of the current technologies for

cooking and to improve designs of mud-stoves already in place in the settlement, is needed. Specific

expertise should be engaged. Existing knowledge and skills in construction and use of improved mud

stoves and increasing awareness to get fuel-efficient stoves in every household should be also

encouraged in Bidi Bidi settlement.

In addition, the following energy savings measures are recommended to further reduce fuelwood

consumption for cooking:

- Use, and production, of heat retention boxes and bags using locally available materials should

be promoted, which can reduce fuel consumption by over 40%.

- Use suitable lids for all cooking task helps to contain the heat so that food cooks faster.

- Soak beans and grind or cut food in smaller pieces. For instance, beans need to be soaked

overnight for 8-14 hours and cooked the next day. Soaking makes beans soft so that they will

cook for a shorter time and using much less fuelwood.

3 http://www.unhcr.org/protect/PROTECTION/438724c42.pdf 4 http://pdf.usaid.gov/pdf_docs/Pnadt959.pdf

33

- Dry fuel before use and process into smaller pieces. Using dry wood would increase cooking

energy efficiency, reducing the amount required for cooking with the side benefit of reducing

harmful smoke emissions.

- Share cooking facilities between families.

- The use of traditional clay cooking pot, although more delicate, absorbs and contains heat

longer than a metal pot and once it gets hot the clay pot requires lesser fuel to keep it going

compared to the metal one.

Tree planting and improved forest management

There is a clear risk of high levels of deforestation and land degradation due to the increased woodfuel

demand due to the sudden influx of refugees in Bidi Bidi settlement. It is a priority to improve the

management of existing forests and other woodlands as well as planting trees to increase the woody

biomass production in the settlement and host communities land. Planning, implementing and

monitoring a forest management plan, in consultation with the local and national authorities as well

as with host and refugee communities, needs urgent attention. In addition, forestry and

environmental expertise should be engaged to guide the planning and implementation process and

build capacity and knowledge for longer-term success of the project.

A forest management plan for Bidi Bidi should aim to support the energy needs of refugee and host

communities and at the same time to reduce the environmental and social impacts as result of

overexploitation of the natural resources, deforestation and forest degradation.

Fast growing tree species with short rotation coppice management should be preferred for an early

fuelwood collection but also multipurpose species can boost motivation of the population in taking

care of trees also for other socio-economic benefits and non-wood forest products such as provision

of fruits and fodder, soil conservation, soil fertility, fencing, etc.

Afforestation, reforestation and ecosystem restoration/rehabilitation interventions should be carried

out in a participatory manner with leaders in the different communities giving a particular attention

to the interlinkages between the environmental and different sociocultural aspects. The

establishment of new tree plantations either for fuelwood production or land

restoration/rehabilitation requires the development of a detailed forest management plan including

key elements such as the selection of tree species, planting material, planting schemes, timeframes,

tree nursery management, suitable land in addition to costs of all stages of the programme, tenure

and legal issues, water supply, fertility, human resources, protection of the site, transplanting, tending

operations, harvesting schemes and closed coordination with local government and other actors

engaged in similar programmes in the region.

“Fuelwood-for-work” programmes could be linked to the ongoing UNHCR “cash-for-work”

programmes similar to WFP “food-for-work” programmes. FAO has decades of experiences in

promoting and monitoring forest and tree resource assessment, sustainable energy systems and

enhancement of ecosystem services. In particular, afforestation, reforestation, ecosystem restoration,

agroforestry systems and promotion of improved energy consumption means can contribute to

addressing environmental degradation, deliver integrated solutions and enhancement of resilience

and adaptation.

34

According to the results from the Scenario with tree plantations and improved stoves, to compensate

the annual biomass loss as result of woodfuel demand, between 9 368 and 12 178 hectares would

need to be planted with fast growing species that corresponds to 12-15% of the total land area of Bidi

Bidi settlement.

The focus group discussions revealed that both the refugees and host communities are interested in

planting trees if technical and financial support is provided. Refugees showed interest in tree planting

even through a fuelwood-for-work approach for planting and monitoring trees. They had some

concerns for other tree species citing constraints of termites, livestock destruction, lack of land, poor

soils in the settlement and water stress area. Fuelwood-for-work seems a welcome and feasible idea

for district authorities, UNHCR staff and the refugees themselves.

Gender equity

Like in the host communities there is a clear marked existence of gender norms and inequities that

disproportionately affect women in the refugee settlement. A gender sensitive approach that

addresses existing inequities but at the same time minimizes their potential for negative effects on

programmes work is needed. Women should also be involved in tree plantation programmes through

the selection of trees planted, based on their specific needs. This could eventually be coupled with

training in alternative livelihood strategies.

Conflict sensitivity, resolution and peace building

Conflict sensitivity should be integrated in all programmes of government and humanitarian agencies

and in community structures. In addition conflict resolution and peace building activities should be

promoted among refugees and host communities. Considering that the refugee population are in Bidi

Bidi as a result of conflict, then peace, peaceful co-existence, and reduced tensions between the host

and refugees communities over natural resources are prerequisites for the two communities to

receive and appreciate benefits from the different humanitarian programmes. Peace-building and

conflict sensitivity and tracking should be a significant part of planning, implementation and

monitoring of all agencies working together with the OPM and information could be reported on

during the regular meetings.

Monitoring

The methodology developed can be regularly reproduced for monitoring the woodfuel supply and

demand in refugee-hosting areas. By using free and public satellite imagery the continuous monitoring

of biomass stock changes can be implemented using open source solutions that can be accessed

through the SEPAL platform at https://sepal.io and should be used to regularly update the findings

from this report.

35

5. Conclusions

The refugees in Bidi Bidi settlement as it is are living in a status of acute emergency and are exposed

to a wide array of vulnerabilities. With their livelihoods disrupted, refugees have no source of income,

lack secure access to fuel and energy for cooking which sets off other vulnerabilities like compromised

nutrition, food security, health and protection to the person. The wider community meanwhile is

affected by issues of tension/conflict and eventually irreversible environmental degradation and

climate change. Practical measures need to be undertaken to urgently reduce some of these impacts

through careful participatory planning, implementation and monitoring.

The resources around the area are scarce and under obvious pressure and could hence, in a

hypothetical worst case scenario, supply the refugees for 3 years. This would however be at the cost

of a full depletion of the area in terms of aboveground biomass, representing an issue that needs to

be addressed as soon as possible. Continuous monitoring of these resources will be needed in the

coming months to decision making on measures to be taken to mitigate the pressure.

The likelihood is that a combination of interventions will be needed, effective immediately, to reduce

the high woodfuel demand from the refugees of Bidi Bidi settlement and to support a sustainable

woodfuel supply. These include:

i) introducing fuel-efficient cookstoves and promoting energy savings measures at the

household level;

ii) establishing multipurpose tree planting programmes for building resilience and creating

opportunities for sustainable livelihoods;

iii) developing community based forest resources management plans;

iv) promoting woodfuel alternatives (e.g. agricultural residues, LPG).

Such interventions would also have the added value of contributing to climate change mitigation

through carbon sequestration and reduction of CO2 emissions. The carbon sequestrated by forest

plantations and the emission reductions from improved cookstoves could also provide benefits

through carbon credits.

36

6. References and further reading

Chave, J., Rejou-Mechain, M., Burquez, A., Chidumayo, E., Colgan, M. S., Delitti, W. B. C., Duque, A., Eid, T., Fearnside, P. M., Goodman, R. C., Henry, M., Martinez-Yrizar, A., Mugasha, W. A., Muller- Landau, H. C., Mencuccini, M., Nelson, B. W., Ngomanda, A., Nogueira, E. M., Ortiz-Malavassi, E., Pelissier, R., Ploton, P., Ryan, C. M., Saldarriaga, J. G., Vieilledent, G. (2014). Improved allometric models to estimate the aboveground biomass of tropical trees. Global Change Biology, 20(10), 3177- 3190. FAOSTAT 2015. Data portal. FAO, Rome, Italy. Available at http://www.fao.org/faostat/en/#home FAO 2015. FRA report, 2015: Global Forest Resources Assessment 2015. Available at http://www.fao.org/3/a-i4782e.pdf FAO/UNHCR 2016. Assessing woodfuel supply and demand in displacement settings, by D’Annunzio, R., Gianvenuti, A., Henry, M., Thulstrup, A. Rome, Italy. Available at http://www.fao.org/3/a-i5762e.pdf Forest Department, Ministry of Water Lands and Environment (2002): National Biomass Study –Technical Report, 113p. GSARS (forthcoming). How to Include a Woodfuel Supplementary Module into Existing Surveys and Derive Woodfuel Indicators. A. Borlizzi, Technical Report no. 3 of the project “Developing a Woodfuel Survey Module for Incorporation into Existing Household Surveys and Censuses in Developing Countries”. Rome, FAO. Gunning, R. 2014. The current state of sustainable energy provision for displaced populations: an analysis. London, Royal Institute of International Affairs, Chatham House. Joshi, N., Baumann, M., Ehammer, A., Fensholt, R., Grogan, K., Hostert, P., Rudbeck Jepsen, M., Kuemmerle, T., Meyfroidt, P., Mitchard, E., Reiche, J., Ryan, C.M., Waske, B. 2016. A Review of the Application of Optical and Radar Remote Sensing Data Fusion to Land Use Mapping and Monitoring. Remote Sensing, 8(1), 70. UBOS 2014. Uganda Population and Housing Census 2014. UNHCR/IUCN 2005. Forest Management in Refugee and Returnee Situations: A Handbook of Sound Practices. Geneva, Switzerland. Available at http://www.unhcr.org/protect/PROTECTION/438724c42.pdf USAID. 2010. Fuel-efficient stove programs in humanitarian settings: an implementer’s toolkit. Washington DC, United States Agency for International Development. Available at http://pdf.usaid.gov/pdf_docs/Pnadt959.pdf Useful Trees and Shrubs for Uganda. Identification and Management for Agricultural and Pastoral Communities. By A. B. Katende, Ann Birnie and Bo Tengnas - Kamapala and Nairobi: Regional Soil Conservation Units (RSCU), Swedish International Development Cooperation Agency (Sida), 1995. Available at http://www.worldagroforestry.org/downloads/Publications/PDFS/b09383.pdf

37

7. Annex

Table A1: Tree species list and number of recordings from the field inventory per LULC class.

Tree species

Closed woodland

Open woodland

Bushland and grassland

Cultivated land

Total

Common uses5

Remarks

Acacia hockii 40 51 11 0 102 Firewood, charcoal, medicine (roots), ropes (bark), fencing (dry branches).

The root is used to treat cough.

Acacia macrothyrsa

0 0 0 0 0

Acacia polyacantha

1 2 0 0 3 Firewood, timber (roofing, tables, matches, boxes, carving, musical instruments), live fence.

Acacia sieberiana

1 5 0 0 6 Firewood, charcoal, timber, fodder (leaves), tool handles, N-fixation, gum, fences.

The foliage and fruit are good fodder and particularly important in the dry areas of northern Uganda. Often left in fields for shade.

Antiaris toxicaria

1 0 0 0 1 Timber (veneer, beer canoes), medicine (leaves, roots), bark cloth.

The root or leaves are used to treat mental illnesses. Plant individual trees for shade, as avenue trees or as a pure stand.

Balanites aegyptiaca

0 1 0 0 1 Firewood, charcoal, poles, timber (furniture), utensils, tool handles, food (fruit, leaves), medicine (roots, bark,

An important species for dry areas as it produces fruit in very dry seasons. Improved varieties exist in India. The wood is termite

5 http://www.worldagroforestry.org/products/switchboard/index.php/name_like/Acacia/

38

fruit), mulch, shade, windbreak, gum, fencing (branches), oil (fruit), fish poison.

resistant. The edible fruits are rich in saturated fatty acids which are used as cooking oil. The leaves are eaten by populations in Adjumani and parts of Teso. https://www.researchgate.net/publication/259921775_Balanites_Aegyptiaca_L_A_Multipurpose_Fruit_Tree_in_Savanna_Zone_Of_Western_Sudan

Borassus aethiopum

0 0 0 1 1 Firewood, charcoal, poles, timber (roofing, door frames, etc.), tool handles, bee hives, food (fruit, seeds, young seedlings), palm wine (sap of flower shoots), medicine (roots, flowers, oil), fodder (fruit, young leaves), thatch, fibre (leaves), baskets, mats (leaf stalks, leaves), oil (fruit, pulp), brooms, drums.

Elephants eat the fruits, thus distributing the tree. The wood is hard and resistant to termites and fungi. Populations in the Acholi area eat the first root shoot from the seed. The fruit is also edible, tasting like peach.

Borassus egyptian

0 2 0 0 2

Bridelia scleroneura

0 7 0 0 7 Medicinal, the wood makes excellent fuel, termite resistant.

Butyrospermum paradoxum

0 1 0 0 1 Firewood, charcoal, timber, food (seed), fodder, oil (cooking, soap, candles), shade, ornamental.

The shea – a vital source of cooking and “smearing oil” (from the kernel) across northern Uganda, although more in the Acholi/Langi areas than West Nile. The fruit is also edible.Tree survives annual fires and the wood is hard and termite resistant and much sought after for making charcoal.

Celtis africana 2 0 0 0 2 Firewood, charcoal, timber (local construction, farm tools), tool handles, fodder (leaves), shade.

The timber rots and splits easily, but it is very strong and with proper seasoning useful for tool handles and building. Leaves are browsed by animals, including cattle, and the leaves and fruits are important in the diet of black and white colobus monkeys. Does not compete with crops

39

since it has a light shade. It is a tree quite suitable for parks and avenues.

Combretum collinum

20 29 9 9 67 Firewood, charcoal, medicine (leaves, roots), bee forage, fencing (cut branches), firebreaks.

Roots are used to treat diarrhoea and vomiting. Makes very good charcoal. Flowers produce good nectar for honey. Suitable for savannah firebreaks. The hard, durable wood burns well but the living tree is tolerant of grass fires.

Combretum fragrans

1 20 7 0 28

Combretum molle

3 10 3 0 16 Firewood, charcoal, poles, posts, timber (construction), tool handles, medicine (roots, leaves, bark), bee forage, mulch.

There are three East African varieties. The hard yellow wood is useful for tools and burns well giving intense heat. The wood is moderately termite resistant. Medicine from the roots has been used for treating hookworm, snake bite, stomach pains, fever, dysentery and leprosy. The charcoal from this species is highly appreciated in Uganda and its planting should be encouraged.

Combretum sp. 0 2 3 0 5

Dichrostachys cinerea

1 0 0 0 1 Firewood, charcoal, poles, posts, tool handles, medicine (leaves, roots), fodder (leaves, pods), bee forage, nitrogen fixation, soil conservation, fibre (bark), live fence, dry fence.

The tree is not planted near houses because it is very thorny. It can be an aggressive weed, has vigorous root suckers and can form a dense thicket. The timber is very heavy and hard but of quite small size. Pounded roots and leaves are used to treat epilepsy. Suitable for planting on degraded soils. It can be an indicator of overgrazing in low rainfall areas.

Dichrostachys sp.

0 2 0 0 2 Firewood, charcoal, poles, posts, tool handles, medicine (leaves, roots), fodder (leaves, pods), bee forage, nitrogen fixation, soil conservation, fibre (bark), live fence, dry fence.

The tree is not planted near houses because it is very thorny. It can be an aggressive weed, has vigorous root suckers and can form a dense thicket. The timber is very heavy and hard but of quite small size. Pounded roots

40

and leaves are used to treat epilepsy. Suitable for planting on degraded soils. It can be an indicator of overgrazing in low rainfall areas.

Ficus natalensis 1 0 0 0 1 Medicine (leaves), shade, live fence, bark cloth.

Barkcloth used to be made from this tree throughout Uganda. The leaves are used to treat dysentery and sore throats. The tree is also grown as a live fence around homes and at a wide spacing for shade in coffee, cocoa and banana plantations. It is emerging as fodder for zero-grazed cows.

Flueggea virosa 1 0 0 0 1 In parts of Africa, the slender branches are used to make fish traps. The small fruit is sweet and eaten by people, animals and birds when ripe. The roots and fruits are believed to be an effective snakebite remedy. Roots of this plant are also used in some African communities as contraceptives and for the treatment of syphilis, gonorrhoea, rheumatism, sterility, rashes, and an infusion of the root is taken to relieve malaria. The bark is believed to provide a treatment for diarrhoea and pneumonia. http://pza.sanbi.org/flueggea-virosa

Gardenia sp. 0 1 0 0 1

Grewia bicolor 5 0 0 0 5 Firewood, charcoal, timber, tool handles, carving (clubs, javelins, walking sticks), medicine (roots, bark), fodder (leaves, fruit).

Twigs from the tree are used by water diviners to locate underground water. The fruit is edible, sweet and remarkably high in iron. It is especially important for pregnant woman.

Grewia mollis 1 33 0 8 42 Firewood, charcoal, timber, walking sticks, fodder (leaves and fruits), fibre (strings from bark).

Fibres under the bark are made into string for general use and in constructing granaries. The heartwood makes excellent walking sticks and the Acholi use the wood to make spears.

41

Harrisonia abysinica

4 0 0 0 4

Hymenocardia acida

0 1 0 0 1 Firewood, charcoal, poles, medicine (bark, leaves), bee forage, ornamental, fibre (inner bark), oil (flowers, seed), soap (roots), brooms (from saplings).

A beautiful flowering tree with potential as an ornamental in parks and gardens. The wood is pale yellow and has been used for bows. The fibres have been used for fish nets, bead strings and thread to sew barkcloth. The roots contain ethylsalicylate and a saponin making them highly poisonous. An antidote for snakebite and a cough mixture are prepared from leaves, an abortifacient from powdered bark and a root infusion relieves toothache. But great care must be taken when using these medicinal substances.

Khaya senegalensis (Mahogany)

0 3 0 0 3 Firewood, charcoal, timber (heavy construction), soil conservation and improvement.

Khaya senegalensis is a hardier species than other Khaya. It can grow in open savannah as well as in forests and is suitable for Northern and North Western Regions. The timber, which is harder and heavier than that of the other Khaya, can be used for construction such as bridges and for railway sleepers. It is also more termite resistant.

Lannea barteri 0 6 1 0 7 Firewood, charcoal, utensils (durable mortars), live fence.

A species that should be encouraged to improve fuelwood supplies. All Lannea are very fire resistant and coppice easily.

Lannea schweinfurthii

1 0 0 0 1 Firewood, charcoal, timber (stools, chairs, mortars), food (fruit), medicine (leaves, bark, roots), fodder (leaves).

The species is resistant to fire. The fibrous roots, like red-brown wool, have been used for stuffing pillows. Young branches are very flexible and the white wood is soft and light.

Lannea sp. 0 1 0 0 1 Firewood, charcoal, utensils (durable mortars), live fence.

A species that should be encouraged to improve fuelwood supplies. All Lannea are very fire resistant and coppice easily.

42

Lonchocarpus laxiflorus

0 2 0 0 2

Maytenus ovata 1 13 0 0 14

Maytenus sp. 0 0 0 0 0 Firewood, charcoal, timber (local construction), farm tools, medicine (roots), live fence, ornamental.

The wood is red and heavy. The trees grow easily from seed. Raise as a pure stand.

Piliostigma thorningii

0 0 0 1 1 Firewood, charcoal, poles, timber (houses), food (pods), drink (leaves, pods), fodder (pods, shoots), bee forage, medicine (leaves, bark, roots, pods), mulch, soil conservation, ornamental, N-fixation, tannin, dye (pods, seeds, bark, roots), rope (bark, root fibres).

A good tree that can be grown mixed with Annona spp., Grewia spp., and Combretum spp. Competes very little with maize if left in fields and pollarded to reduce shade. The pulp surrounding the seeds is edible and under famine conditions leaves, crushed green pods and seeds have been eaten. Pods and seeds give a blue dye and roasted seeds a black dye.

Pseudocedrela kotschyi

0 5 0 0 5 Firewood, charcoal, timber (joinery), utensils (mortars, etc.), poles, shade, ornamental.

An easy-to-grow species for the semi-arid savannahs. The timber is attractive, resembling mahogany but harder and heavier, and is suitable for high-class joinery. Mortars are made from the trunks. Can be grown in pure stands, for shade, as an avenue tree or interplanted with mangoes, Cassia spp. and cashew nut.

Rhus natalensis 4 1 0 0 5 Firewood, timber (utensils), tool handles, food (fruit), medicine (roots, bark, leaves), fodder, oil (seed), live fence.

A useful tree for semi-arid areas as it is drought resistant. The wood is heavy, hard and very durable. The seed contains up to 60% of a non-drying oil suitable for soap and lubrication. It has also been used as body and hair oil and for softening leather.

Senna sp. 0 0 0 2 2

Sterculia setigera

0 5 0 0 5 Fibre (bark), bee hives (coppice shoots).

Once the bark is removed, the stem dies. The bark is used to make baskets.

Strychnos spinosa

2 1 0 0 3 Firewood, charcoal, timber (furniture, boxes), fodder (leaves), food (fruit),

Although the fruit is edible, seeds are toxic and unripe fruit may be also. The fruit are often

43

medicine (fruit, leaves, bark, roots), musical instruments (dry fruit shell).

eaten by wild animals. The wood is straight- grained and planes well but has not been much used for carpentry in Uganda.

Tamarindus indica

7 0 0 0 7 Firewood, charcoal, poles, timber (furniture, boats, general purposes), food (pulp for drink, fruit, spice), fodder (leaves, fruit), medicine (bark, leaves, roots, fruit), mulch, N- fixation, shade, ornamental, windbreak, tannin (bark).

The dark brown heartwood is hard and heavy, well grained and easy to polish. The pulp is rich in vitamin C. It is recommended for homestead planting and along river banks.

Terminalia glaucens

0 0 0 1 1 Firewood, charcoal, bee hives. Produces excellent charcoal. The timber is yellow-brown, hard, durable and tough and works well.

Unknown 5 0 0 0 5

Unknown Lannea

1 0 0 0 1 Firewood, charcoal, utensils (durable mortars), live fence.

A species that should be encouraged to improve fuelwood supplies. All Lannea are very fire resistant and coppice easily.

Ximenia americana

0 3 0 0 3 Firewood, timber (utensils), tool handles, food (fruit), medicine (roots, bark, leaves), fodder, oil (seed), live fence.

A useful tree for semi-arid areas as it is drought resistant. The wood is heavy, hard and very durable. The seed contains up to 60% of a non-drying oil suitable for soap and lubrication. It has also been used as body and hair oil and for softening leather.

Ziziphus abyssinica

7 0 0 0 7

Firewood, charcoal, timber (furniture, interior work, carving), poles, food (fruit pulp and seed), live fence, bee forage, dye (bark).

The spiny branches make this plant useful as a protective live fence. Leaves and fruit are used elsewhere as fodder. The bark yields a cinnamon-coloured dye. The timber is heavy, hard and resistant to termites and borers.

44

Table A2: Shrub species list and number of recordings from the field inventory per LULC class.

Shrub species

Closed woodland

Open woodland

Bushland and grassland

Cultivated land

Total

Alchornea laxiflora

0 0 0 1 1

Flueggea virosa 9 1 0 5 15 Harrisonia abysinica

6 0 0 0 6

Rhus natalensis 3 0 0 0 3

Unknown 4 10 2 0 16

22 11 2 6 41