ABSTRACT
Title of dissertation: AGRICULTURAL LAND USE,
DROUGHT IMPACTS AND VULNERABILITY:
A REGIONAL CASE STUDY FOR
KARAMOJA, UGANDA
Catherine Lilian Nakalembe,
Doctor of Philosophy, 2017
Dissertation directed by: Professor Christopher O. Justice
Department of Geographical Sciences
The increasing frequency of extreme climate events brings into question the
sustainability of agriculture in marginal lands, especially those already experiencing
drought such as the Karamoja region in northeastern Uganda. A significant amount of
research often qualitative has been conducted documenting drought and its impact on
Karamoja. Taking a mixed methods approach, this study combined remotely-sensed
satellite data, national agricultural surveys, census, and field data to expand on
empirical knowledge on agricultural drought, land use and human perceptions of
drought necessary for comprehensive drought forecasting, monitoring, and
management. Results from this study showed that Karamoja is at least twice more
vulnerable to drought than any other region in Uganda. This is because of its very low
adaptive capacity in part due to high poverty rates and a higher dependency on the
natural environment for livelihood. Analysis of satellite data quantified a 229 percent
increase in cropland area in Karamoja between 2000 and 2011/12, driven largely by
agricultural development programs. Underlying forces (e.g., cropland expansion
programs and controlled grazing) originating from land use policy and development
programs, more than proximate causes (direct local level actions) remain the major
drivers of this expansion. Although the cultivated area has dramatically increased,
there is no quantifiable overall increase in yield or per-capita production as evidenced
by the recurrent poor food security. This status quo, (poor yields and dependence on
food aid) is likely to continue as more land is put to crop cultivation by poor
households and meager investments are made in livestock-based livelihood
opportunities.
The cropland area mask developed in this research facilitated the characterization
of drought within agricultural areas. The drought information developed by this study
is spatially and temporally explicit, showing differences in severity between years and
between districts. Overall Abim District showed the least variation and is the least
impacted while, Moroto District had the highest inter-annual variability and was often
the most severely impacted. This research presents an approach to predict the number
of people who would require food aid during the lean season in Karamoja (December
to March) within a reasonable margin of error (less than 10%) at the peak of the
growing season (August/September), although the need for more extensive testing is
recognized. The method takes advantage of readily available satellite data and can
contribute to planning for a timely and appropriate response.
A case study of farmer’s perceptions of drought in Moroto District found that
many farmers feel helpless and have no control of their future. For the majority of
farmers in the district, past experiences of drought do not necessarily impact on future
expectations of drought and many have no long-term adjustment plans. Quite often
the majority of the population depends on emergency food assistance, building a
culture of dependency. The analysis indicates that factors such as; conflict (insecurity)
and interventions by government and international agencies intermingle with culture to
have a profound direct influence on farmers’ perception of drought amongst
communities in Moroto district. This research shows that satellite data can provide the
much-needed information to fill the gaps that inhibit long-term drought monitoring, at
a significantly lower cost than traditional climate station-based monitoring in data
scarce regions like Karamoja. It also points to a way forward for proactive assessment,
planning, and response..
AGRICULTURAL LAND USE, DROUGHT IMPACTS AND
MITIGATION: A REGIONAL CASE STUDY FOR KARAMOJA,
UGANDA
by
Catherine Lilian Nakalembe
Dissertation submitted to the Faculty of the Graduate School of the
University of Maryland, College Park in partial fulfillment
of the requirements for the degree of
Doctor of Philosophy
2017
Advisory Committee:
Professor Christopher O. Justice, Chair/Advisor
Research Assistant Professor. Jan Dempewolf
Professor Christopher Jarzynski
Associate Professor. Julie Silva
Research Professor Eric Vermote
c© Copyright by
Catherine Lilian Nakalembe
2017
Foreword
The research presented in this dissertation has been submitted to peer-reviewed
journals, with myself as first author and others (including some members of my
dissertation committee) as co-authors. The research presented in Chapter 3 has been
published by The Land Use Policy Journal in March 2017 (Nakalembe et al., 2017) and
Chapter 4 is currently under review by the Natural Hazards Journal (Nakalembe,
2017). These two chapters form the basis of the Disaster Risk Financing Program
(DRF) that aims to enable timely scaling up of Labour Intensive Public Works in case
of drought under the Third Northern Uganda Social Action Fund (NUSAF 3). NUSAF
3 aims to empower communities in Northern Uganda by enhancing their capacity to
systematically identify, prioritize, and plan for their needs and implement sustainable
development initiatives that improve socio-economic services and opportunities. The
DRF mechanism relies on satellite information for early warning to ensure additional
funds are available in time to expand the public works program for affected
communities. In addition, this work informed the first ever government response to
drought in Karamoja based on a report I wrote on food security-based data and
methods presented in Chapter 3 and 4. This has also facilitated continued drought
monitoring using satellite data by the Office of the Prime Minister. Chapters 2 and 5
are combined to form the third publication of this dissertation that will be submitted.
ii
Dedication
I dedicate this work to my mother Rita Nanono-Nalongo, your hard work,
dedication, and inspiration got me here, and to Paul Lokeris of Natopojo, Moroto, may
all your dreams and wishes come true.
iii
Acknowledgments
I would like to express the deepest appreciation to my Advisor and dissertation
committee chair, Dr. Chris Justice. Dr. Chris has the attitude and the substance of a
genius. Without his vision and persistent support, I would have missed out on all the
practical learning I have had over the course of my studies. He has allowed me to work
independently and provided me opportunities and challenges only the best education
can offer. For this, I will forever be grateful.
I would like to thank my committee members, Dr. Jan Dempewolf, Dr. Julie
Silva, Dr. Eric Vermote, and Dr. Christopher Jarzynski whose work and research
interests are so diverse; they demonstrated to me the true value of a good education. I
have come to appreciate and understand the value of working with the best. My
colleagues, Inbal Becker-Reshef, Christina Justice, Alyssa Witchcraft, Mike Humber
and Brian Barker from the GLAM team; thanks for letting me be part of one of the
best teams.
I would like to thank the Commissioner of the Department of Disaster
Preparedness, Relief and Management Mr. Martin Owor for allowing me a seat at the
table. Without his visionary approach, I would have never had the opportunity to be
involved in the Disaster Risk Financing Program in Karamoja.
I cannot thank you enough John Olinga (Moroto District Agricultural Officer);
you taught me that no matter the situation, staying calm is being brave. I always
looked forward to field work in Karamoja knowing that you will be there with me and
for me. I enjoyed every story you shared that opened my heart to the people of
iv
Karamoja. Because of you, I was able to meet Paul Lokeris. My friends in Moroto
Janaan Edonu and Helen Maraka thank you for walking those fields with me,
especially after the rain. I always looked forward to our regroup at the end of each day.
To my family, the hardships, struggles, and triumphs with you make me who I
am. I have grown knowing love and hard work. I thank our mother, the true leader of
our band, one who welcomes everyone to our home, always giving second chances. My
dad Steven Busulwa a true source of inspiration. My sisters; Annet Nakamya for
always believing in me and supporting me every step of the way. Rwiza Nakiyingi for
always praying for me, and Norah Nassimbwa for always loving me. And lastly, to my
husband Sebastian, your patience, kindness, and love keep me strong and alive.
My graduate research was supported by funding from the NASA Land Cover and
Land Use Change Program (LCLUC) that also enabled access to tools and data
including very high-resolution satellite data. Fieldwork in 2014 was possible through
the Spurring a Transformation in Agriculture Through Remote Sensing (STARS)
project.
v

Contents
Foreword ii
Dedication iii
Acknowledgements iv
Table of Contents vi
List of Tables ix
List of Figures xi
List of Abbreviations xii
1 Introduction 1
1.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Research Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
1.3 Dissertation Organization . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2 Mapping drought vulnerability in Uganda 11
2.1 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
2.2 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
2.3 Study area . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.4 Data and methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
2.4.1 Assessing drought vulnerability in Uganda . . . . . . . . . . . . . 20
2.4.2 Assessing exposure . . . . . . . . . . . . . . . . . . . . . . . . . . 21
2.4.3 Assessing sensitivity . . . . . . . . . . . . . . . . . . . . . . . . . 22
2.4.4 Assessing adaptive capacity to cope with drought . . . . . . . . . 24
2.5 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
2.5.1 Exposure to drought in Uganda . . . . . . . . . . . . . . . . . . . 25
2.5.2 Sensitivity to agricultural drought in Uganda . . . . . . . . . . . 25
2.5.3 Sub-regional adaptive capacity to cope with drought impacts in
Uganda . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
2.5.4 Vulnerability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
2.6 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
3 Agricultural Land Use Change in Karamoja Region, Uganda 37
3.1 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
3.2 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
vii
3.3 Study Area: The Karamoja Region of Northeastern Uganda . . . . . . . 40
3.4 Land use history . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
3.5 Agricultural land use in Moroto District Karamoja . . . . . . . . . . . . 51
3.6 Estimating cropland expansion in Karamoja with remote sensing data . . 55
3.7 Cropland expansion in Karamoja . . . . . . . . . . . . . . . . . . . . . . 60
3.8 Summary of factors contributing to cropland expansion in Karamoja . . 62
3.9 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
4 Characterizing Agricultural Drought in the Karamoja sub-region of Uganda
with Meteorological and Satellite-Based Indices 69
4.1 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
4.2 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 70
4.2.1 Defining and characterizing drought . . . . . . . . . . . . . . . . . 74
4.2.2 Study Area . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
4.3 Data and Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 79
4.3.1 Standardized Precipitation Index . . . . . . . . . . . . . . . . . . 79
4.3.2 Normalized Difference Vegetation Index . . . . . . . . . . . . . . . 81
4.3.3 Agricultural land use data . . . . . . . . . . . . . . . . . . . . . . 82
4.3.4 NDVI and SPI Relationships . . . . . . . . . . . . . . . . . . . . . 86
4.3.5 Investigating the relationships between NDVI anomaly and food
insecurity in Karamoja . . . . . . . . . . . . . . . . . . . . . . . . 87
4.4 Results and discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
4.4.1 Correlation between rainfall and crop/vegetation conditions . . . 91
4.4.2 Drought identification from rainfall (SPI-3) data . . . . . . . . . . 96
4.4.3 Spatial and temporal patterns of agricultural drought 2000 to 2012
from NDVI data . . . . . . . . . . . . . . . . . . . . . . . . . . . 100
4.4.4 Relationship between NDVI and number of people requiring food
aid . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107
4.5 Concluding remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
5 Human Perceptions of Drought in Karamoja Uganda: a Moroto District
Case Study 115
5.1 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
5.2 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117
5.3 Materials and Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
5.3.1 Study Area . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
5.3.2 Data Collection and Analysis . . . . . . . . . . . . . . . . . . . . 124
5.4 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
5.4.1 Drought experience and memory . . . . . . . . . . . . . . . . . . 127
5.4.2 Farmer definitions of drought . . . . . . . . . . . . . . . . . . . . 134
5.4.3 Expectations of drought . . . . . . . . . . . . . . . . . . . . . . . 136
5.4.4 Coping with drought in Moroto . . . . . . . . . . . . . . . . . . . 138
5.5 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
5.6 Conclusion and policy recommendations . . . . . . . . . . . . . . . . . . 143
viii
6 Conclusion 147
6.1 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
6.2 Future Research . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 153
6.3 Policy Implications and Next Steps . . . . . . . . . . . . . . . . . . . . . 155
Appendix 159
Bibliography 163
ix

List of Tables
2.1 Data use to assess exposure, sensitivity and adaptive capacity . . . . . . 21
2.2 Drought categories and possible impacts adapted from United States
Drought Monitor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
2.3 Sub-regional exposure to drought in Uganda (where D0-Mild, D1-Moderate,
D2-Severe, D3-Extreme, and D4-Exceptional Drought) . . . . . . . . . . 27
2.4 Sub-regional poverty rates . . . . . . . . . . . . . . . . . . . . . . . . . . 28
2.5 Land ownership by sub-region shown as percent of sample from the Uganda
National Panel Survey of 2014 . . . . . . . . . . . . . . . . . . . . . . . . 29
2.6 Summary of indicators of adaptive capacity used in this study from the
Uganda National Panel Survey of 2014 . . . . . . . . . . . . . . . . . . . 31
2.7 Summary of indicators of adaptive capacity used in this study from the
Uganda National Panel Survey of 2014 . . . . . . . . . . . . . . . . . . . 31
3.1 Moroto District production data compared to Northern Uganda (which
includes Karamoja region) and national production as reported in the
2008/9 census of Agriculture (Uganda Bureau of Statistics, 2010b) . . . . 53
3.2 Examples of land cover products available for Africa, for details see Fritz
et al. (2010) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
3.3 Karamoja cropland area in hectares and percent change estimates between
2000 and 2011/12 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
3.4 Estimate of percent area under cultivation in Karamoja . . . . . . . . . . 62
4.1 Datasets used in this analysis include rainfall data from Kotido Station
and Karamoja District boundaries subset from the national dataset . . . 81
4.2 NDVI based indices used in the analysis. Each index was calculated from
NDVI maximum value composite in 14-day time-steps from 12-year (2000
to 2012) time series data. . . . . . . . . . . . . . . . . . . . . . . . . . . . 81
4.3 Dummy Variable Assignment . . . . . . . . . . . . . . . . . . . . . . . . 87
4.4 SPI time scales 1, 3, 6, 9 and 12 months and NDVI based indices at
Kotido Station within cropland areas N=48. The highest R2 values are
shown in bold for each NDVI index. Overall ANDVI data had the highest
correlation with SPI-3. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 91
4.5 Regression analysis on ANDVI and SPI-3 Kotido. aF= 22.012, p-value<
0.0001, R2=0.324, Adjusted R2=0.309, bF= 6.457, p-value< 0.0001, R2=0.375,
Adjusted R2=0.317 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94
xi
4.6 Pearson Correlation Coefficients (r) between ANDVI (min, mean and
max) values and food aid need (n=5), and IPC Phase 3 (n=8). Bold
r-values indicate cases with highest correlation coefficient.*Indicate val-
ues significant at the 0.05 . . . . . . . . . . . . . . . . . . . . . . . . . . 108
5.1 Characteristics of the surveyed farmers in Moroto District . . . . . . . . 128
5.2 Main thematic statements from interviews under guiding question 1 . . . 129
5.3 Age and experience of drought by farmers in Moroto District . . . . . . . 133
xii
List of Figures
1.1 Schematic of issues addressed in this thesis . . . . . . . . . . . . . . . . . 8
2.1 Study Area: Uganda . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.2 Population trends in Uganda from 1911 to 2014.*=projected population . 16
2.3 Average monthly rainfall at selected stations. . . . . . . . . . . . . . . . . 17
2.4 Uganda livelihood zones (FEWS NET, 2015) . . . . . . . . . . . . . . . . 18
2.5 A farmer stands in his sorghum field that was destroyed by the 2015
drought in Rupa, sub-country, Moroto District Karamoja (a). Destroyed
maize (b) and beans (c) in Lwengo and Rakai District respectively and
were destroyed by the 2016 drought. . . . . . . . . . . . . . . . . . . . . 19
2.6 Sub-regional exposure to drought in Uganda . . . . . . . . . . . . . . . . 26
2.7 Sub-regional sensitivity to drought in Uganda . . . . . . . . . . . . . . . 28
2.8 Sub-regional adaptive capacity to drought in Uganda based on social eco-
nomic indicators from the Uganda National Panel Survey of 2014 . . . . 30
2.9 Sub-regional vulnerability to agricultural drought in Uganda . . . . . . . 32
2.10 Summary of exposure, sensitivity, adaptive capacity, vulnerability to drought
in Uganda . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33
3.1 Karamoja Livelihood Zones 2015 Source: Integrated Food Security Phase
Classification (IPC) Report . . . . . . . . . . . . . . . . . . . . . . . . . 41
3.2 (a) A 1960 Arial Photo of A Manyatta in Karamoja (Deshler, 1960). Mod-
ern day Karamoja, An aerial photomosaic of a Manyatta of Pupu village,
Pupu parish, Rupa sub-county (b) [Photo taken August 20, 2015] using
an unmanned aerial vehicle (UAV) with permission from Chief Executive
Officer, Moroto District and the Uganda Peoples Defense Forces (UPDF).
Kids stand in front of their home in Kadilakeny, Rupa (c) [Photo taken
January 21, 2016] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
3.3 Land use history 1960 to 2002: Conservation areas dropped from 94.6% in
1965 to 40.8% in 2002 (Rugadya and Kamusiime, 2013). Degazetting fur-
ther negatively impacted pastoralists by curbing access to pasture in the
newly designated wildlife reserves but opened up more land to sedentary
agriculture (Kra¨tli, 2010) . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
3.4 Population dynamics in Karamoja. Source data: MAAIF (2010); Mirzeler
and Young (2000); UBOS (2004, 2009); Uganda Bureau of Statistics
(2010b, 2002); Burns et al. (2013) . . . . . . . . . . . . . . . . . . . . . . 50
xiii
3.5 An aerial image of a perfectly demarcated 0.394 hectares field in Rupa
sub-county, Moroto District (a) and (b) is a photos of sorghum fields in
Moroto District including a field ploughed under the tractor hire scheme
(February 2011) (top left) . . . . . . . . . . . . . . . . . . . . . . . . . . 52
3.6 Spatial details obtainable from different satellite systems. With World-
View 2 data at 2m spatial resolution, individual fields (averaging 1.6
hectares) can be delineated. This is not feasible at lower spatial reso-
lution i.e. using Landsat (30m) and MODIS (250m) data. . . . . . . . . . 57
3.7 Footprints of 185 Worldview 1 and Worldview 2 Images with the least-
percent cloud cover selected from over 1136 images for digitization. Data
Source-Digital Globe: NextView License . . . . . . . . . . . . . . . . . . 58
3.8 Derived 2012 croplands map showing the extent of agricultural land use
in Karamoja . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
3.9 Cropland Area 2000 (SERVIR USAID) (a), New Cropland Area Map
2011/12 (b) and New Croplands area highlighted in (c) . . . . . . . . . . 61
3.10 2016 livestock ownership in Karamoja (World Food Program, 2016) . . . 64
3.11 Vegetable and maize gardens under drip irrigation in Nadiket, Moroto
January 2016 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66
4.1 (a). A failed sorghum field in Rupa Sub-county, 13 August 2015. (b)
A moderately impacted field in Nadunget Sub-county, 14 August 2016.
These two fields are within 10 kilometers distance. . . . . . . . . . . . . . 72
4.2 Location of Karamoja Sub-region, showing Meteorological Stations with
fairly complete records. The rainfall gradient decreasing annual average
rainfall from west to east (Rainfall data source: Hijmans et al. (2005) . . 78
4.3 The 2012 croplands map used in this analysis (Nakalembe et al., 2017).
These data show the extent of agricultural land use in Karamoja and
the mean NDVI values within the masked croplands during the month of
August . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84
4.4 NDVI temporal profile at Kotido Station (2000 to 2012). The rain season
is April to September and NDVI values are moderate (0.4 to 0.5) during
the planting months of April and May. The highest NDVI values (peak
greenness) are during the months of June, July, August and September
(0.55 to 0.6) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85
4.5 IPC assessment summary map for Karamoja (valid November 2015 to
May 2016) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89
4.6 IPC assessment summary map for Uganda (valid September 2009 to Jan-
uary 2010) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
4.7 Relationship between SPI (at different time steps) at Kotido Station and
NDVI anomaly data for crops only data 2001 to 2012. . . . . . . . . . . . 93
4.8 Relationship between SPI-3 at Kotido Station and NDVI anomaly data
for crops only data 2001 to 2012. a) Simple regression, b) regression with
dummy variables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
xiv
4.9 June to September (JJAS) crops only observed ANDVI and predicted
ANDVI using simple regression and regression with dummy variables
time-series data for 12 years at the Kotido Station (2000 to 2012). . . . . 97
4.10 August SPI-3 Values 1960 to 2012. The droughts are categorized based
on SPI values as: mild (0 to -0.99) in yellow, moderate (-1.00 to -1.49) in
beige, severe (-1.50 to -1.99) in orange, and extreme in red, when values
fell below -2.00 (McKee et al., 1993). . . . . . . . . . . . . . . . . . . . . 98
4.11 The June-September NDVI Anomaly crops only time-series data for Abim,
Amudat, Kaabong, Kotido, Nakapiripirit, Napak and Moroto Districts
from 2000 to 2012. The blue arrows highlight the difference in severity
between Moroto and Abim districts. . . . . . . . . . . . . . . . . . . . . . 101
4.12 NDVI Anomaly data for the end of August (2001 to 2012) . . . . . . . . 103
4.13 End of month 16-day NDVI Anomaly maps from May to September for
2002, 2008 and 2009 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
4.14 (a) Observed and projected (from ANDVI) number of people needing food
aid and (b) reported and projected (from ANDVI) number of people in
IPC Phase 3. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 110
5.1 Location of study area in Uganda . . . . . . . . . . . . . . . . . . . . . . 122
5.2 Average rainfall for Moroto District. Rainfall data from CHIRPS Data
and NDVI Data from GLAM East Africa . . . . . . . . . . . . . . . . . . 123
5.3 Elements of drought perceptions (Taylor et al., 1988) extracted from
(Slegers, 2008) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126
5.4 Severity of droughts in Karamoja 1960 to 2012. August SPI-3 Values 1980
to 2012. The droughts are categorized based on SPI values as: abnormally
dry less than 0 to -0.7, moderate (-0.80 to -1.2) in pink, severe (-1.3 to
-1.5) in orange, extreme (-1.6 to -1.9) in brown, and exceptional in red,
when values fell below -2.00 (McKee et al., 1993). . . . . . . . . . . . . . 130
5.5 Average vegetation conditions from NDVI data in Nadunget, Katikekile
and Rupa sub-counties in 2009 compared to long-term average, maximum
and minimum NDVI values between 2000 and 2016. Average conditions
shown similar trends in all three sub-counties with clear deterioration to
the worst on record by July 19 . . . . . . . . . . . . . . . . . . . . . . . . 132
5.6 Rainfall anomaly data for Moroto District shows extreme dry conditions
until the 3rd dekad of August. . . . . . . . . . . . . . . . . . . . . . . . . 136
5.7 Farmer recollection of droughts. . . . . . . . . . . . . . . . . . . . . . . . 138
5.8 Left: A lady reveals the only food she has in storage - dry leaves from
a local tree. Top left: Sun-drying cassava peels. Bottom right: villagers
from Kadilakeny carrying rocks to cover up erosion gullies as part of a
‘Food for Work Program’ in January 2016. . . . . . . . . . . . . . . . . . 140
xv
List of Abbreviations
ACTED Agency for Technical Cooperation and Development
ANDVI Anomaly of Normalized Difference Vegetation Index
AVHRR Advanced Very High Resolution Radiometer
CHIRPS Climate Hazards Group InfraRed Precipitation
DEM Digital Elevation Model
FEWSNET Famine and Early Warning System
GDP Gross Domestic Product
GLAM Global Agriculture Monitoring
GPS Global Positioning System
HDI Human Development Index
IPCC Intergovernmental Panel on Climate Change
IRB Institutional Review Board
ITCZ Inter-Tropical Convergence Zone
JJAS June, July, August, and September
KALIP Karamoja Livelihoods Programme
LRA The Lord’s Resistance Army LRA
LTDR The Long Term Data Record
MAM March, April, May
MERIS Medium Resolution Imaging Spectrometer
MODIS Moderate Resolution Imaging Spectroradiometer
NASA National Aeronautics and Space Administration
NDVI Normalized Difference Vegetation Index
NDWI Normalized Difference Water Index
NEMA National Environment Management Authority
NGA National Geospatial Agency
NUSAF Northern Uganda Social Action Fund
OPM Office of the Prime Minister
PDSI Palmer Drought Severity Index
SON September, October, November
SPI Standardized Precipitation Index
UBOS Uganda Bureau of Statistics
UNDP United Nations Development Programme
UNFAO United Nations Food and Agricultural Organization
UNHS Uganda National Household Survey
UNPS Uganda National Panel Survey
UPDF Uganda Peoples Defense Forces
USAID United States Agency for International Development
WFP World Food Program
CDO Community Development Officers
DAO District Agricultural Officers
UNCST Uganda National Council for Science and Technology
xvi
1 Introduction
1.1 Introduction
Current social-economic conditions in pastoral (agro-pastoral) East Africa bring
into question the effectiveness of aid programs and policies in marginalized and poorly
understood regions. Pastoralists in Africa are considered the continent’s most
vulnerable group. To this day despite continued inflow of food aid and poverty
alleviation programs, extreme vulnerability due to entrenched poverty is vividly
manifested in the pastoralists communities during recurrent droughts that continue to
have severe detrimental impacts (Little et al., 2008; Juma, 2009).
In most arid and semi-arid environments rainfall is a key determinant of the
constraints human societies have to adapt to, and mobility (pastoralism) by far has
proved to be the most effective way of addressing the challenges (Niamir-Fuller, 1999).
Sustainable mobility, however, requires that there are structural and legal mechanisms
that allow for the self-evolution of pastoral communities. This today, would require
government support to protect land rights of such communities, but few if any
governments seek to protect such rights (Moritz, 2003). However, national
1
governments often see pastoralists as a problem and today pastoral communities
around the world are reducing mainly due to advancing agriculture and land enclosure,
that are often government driven (Blench, 2001)
The African human-environment system is complex particularly for pastoral
societies; a muddle in many ways to outsiders. African rangelands have attracted
controversy since the beginning of the colonial era. Colonial leaders strongly believed
and worked against pastoral communities (Mortimore, 1998). To illustrate this
(Gilbert, 2007) quotes, Sir Charles Elliot:
“I cannot admit that wandering tribes have a right to keep other superior
races out of large tracts of land merely because they have acquired the
habit of struggling over more land that they cannot utilize.” Sir Charles
Eliot, Kenya Land Report, 19331
Such narratives, forming a vast literature on pastoral communities remain the
key resource for authors and aid agencies, who often rely on past narratives of
degradation, with little regard to the current economics, ecology and/or production
systems (Blench, 2001).
Understanding a complex system requires more than just climate-biased snap
shot assessments. It necessitates holistic approaches that incorporate the human,
environment and culture (Mortimore, 2005). However, since colonial times, the trend
in pastoral East Africa has been to try and regulate what has been seen as inefficient,
degrading and unhealthy indigenous livestock enterprises (Blench, 2001). This
modernization model greatly undermines indigenous knowledge and has failed
1Kenya Land Commission Report (Nairobi: Government Printer,1933)
2
miserably in the past due to miss-interpretation of the African terrain ecology and
culture (Mortimore, 1998). Population growth and resource scarcity are the premises
of the modernization model. The current trends in climate change, resource scarcity,
and agricultural intensification seem to point to the inevitability of the settlement of
pastoralists.
Pastoral communities have not had the luxury to adapt to various environmental
pressures on their own, they have been forced to adapt and continue to be pressured
and influenced by outsider cultures. Pastoralists have adjusted in ways necessary to
match outside pressure by acquiring more arms, reverting to unsustainable practices
such as tree burning and dependency on food aid. The outcome has been continuous
conflicts that are failing the pastoral systems of sub-Saharan Africa. Cattle raiding
among pastoral groups of Eastern Africa transformed from a reciprocal cultural
activity into uncontrollable, technologically sophisticated and highly violent
dimensions (Simala and Amutabi, 2005). East Africa in the early 1990’s became a
permanent war zone where small arms were common among pastoralist and sometimes
more sophisticated than those owned by state forces (Simala and Amutabi, 2005). The
extent of raids drew formerly non-violent communities to slowly acquire arms for
protection from cattle raids. In Kenya for example, the Sumabru pastoralists began to
arm themselves after repeated raids by the Turkana. This exacerbated the conflict into
a deadly and perilous situation (Simala and Amutabi, 2005).
As the marginal cost of securing scarce resources increases, private needs precede
those of the community. New mechanisms for distributing access and laying claim to
resources through new rules, laws, and institutions are created. Many pastoral areas
3
have been institutionalized without the participation of the pastoralist themselves
(Robinson, 2009). Change has become inevitable as land becomes scarcer with
population growth and more private land acquisition. Pastoralists have had to
diversify into agriculture, business and manual labor but livestock still remain
important to their culture and livelihood (Galvin et al., 2008).
Pastoral societies face more threats to their way of life now than at any previous
time (Fratkin, 2001). Alongside worsening climatic conditions, population growth adds
another dimension to the scale of natural and man-made disasters pastoral
communities face (Fratkin, 2001). Other challenges pastoralists face today include high
losses of herding lands to private farms, ranches, game parks, and urban areas;
increased commercialization of their livestock economy; out-migration to urban areas;
periodic dislocation brought about by drought, famine, and civil unrest (Moritz, 2003;
Niamir-Fuller, 1999; Nakalembe et al., 2017). Despite the enormous pressure from
globalization, pastoral societies continue to be resilient (Gray, 2000; Galvin et al.,
2008).
Though considered a viable alternative to crop-based livelihood as this research
will document, the pastoral-livelihood system is often considered an outdated way of
life (Little et al., 2008; Gartrell, 1985; Kra¨tli, 2010; Powell, 2010). Part of the problem
is that the majority of pastoral East Africa is blamed for constant internal and
external conflicts (Gartrell, 1985; Kington, 2002). Invasion of villages (lands and
resources) is intrinsic to pastoral mobility. However, studies have shown that mobile
pastoral production is less prone to drought. It reduces shocks and those who lead a
pastoral life are considered less vulnerable when compared to other households such as
4
those dependent on crop cultivation in the same areas (Little et al., 2008).
Pastoralists traditionally relied on communal land tenure, which allowed them
access to wide ranges of grazing land. Today, land tenures of most African
governments do not accommodate for communal land ownerships (Robinson, 2009).
The mainstream view is that communal or common property is inappropriate land use
that should be transformed (Ellis and Swift, 1988b,a). For example, since the 1930s
the Machakos Reserve in Kenya has been viewed by conservationists to have every
phase of misuse of land. This misuse has been blamed for soil erosion, deforestation
and for keeping the communities in poverty and misery (Mortimore, 2005; English
et al., 1994). However, various studies agree that communal land tenure has been and
still is the most appropriate form of property ownership in dry lands. Positive linkages
between population growth and environmental management are evident and under the
right condition, resilience is prominent (Galvin et al., 2008; Mortimore, 2005).
The fact remains that many governments still take on the Malthusian crisis
assumptions that communal land equals desertification and destruction (Boateng,
2017). Communal ownership is therefore not recognized as sustainable or appropriate
by many governments and non-pastoral communities (Morton, 2007). Pastoralists are
thus in constant conflict with governments because they do not recognize national or
international boundaries that have over time been imposed on them (Robinson, 2009).
They continue to defy governmental laws, authority, and boundaries. This is
contradictory to what is considered well-functioning common regimes (Robinson, 2009).
Climate and climate change are at the center of pastoral conflict. Harsher and
longer droughts have necessitated more movement in search of water and pasture to
5
avoid loss of livestock. Commoditization of livestock, access to illegal fire arms,
privatization of pastoral lands, and declaration of protected lands by governments
continue to marginalize pastoralists. Environmental degradation reinforces conflicts
and as dry lands are becoming dryer and rainfall more sporadic, migration into other
lands in search of pasture and food is inevitable.
The question that arises is what can governments and other agencies do to limit
the suffering that results from climatic and external pressures (Robinson, 2009).
Societies that undergo externally-driven disruptions and reorganizations often
experience permanent irreversible vicious circles. Food aid often distributed as a
solution to starvation and malnutrition if continuous reduces the need to reorganize
and adapt to the change that is occurring in recipient communities. Kenyan Gabra
pastoralists are a good example of a society that has been crippled by food aid
(Robinson, 2009).
This thesis aims to address this question, focused on one of the most
impoverished regions of Africa. The Karamoja Region in northeastern Uganda like
other pastoral regions in Sub-Saharan Africa faces several natural and man-made
disasters. Among these and of special interest for this research is drought. Cyclical
droughts and erratic rainfall continue to affect crop production and pasture, and
directly affect the livelihoods of the population. Subsistence agro-pastoralism remains a
major source of livelihood for the Karamajong population that is largely malnourished
due to constant famine (Stark, 2011; Jones, 2011). In recent years, droughts in
Karamoja have anecdotally been more frequent and severe. Farmers in Karamoja bear
the brunt of the drought impact. Extreme drought affects their livelihoods, resulting in
6
famine, migration, and loss of life. Despite the continued flow of aid into the region,
extreme climatic conditions, crop failure, and political instability have kept the region
in dire poverty and exceedingly vulnerable to drought and its impact. (Stark, 2011;
Baker, 1974; Gartrell, 1985). Communities in Karamoja like every other society adjust
their practices in response to actual or expected impacts of natural disasters to
minimize adverse effects and sustain their livelihoods. Coping methods adopted in
Karamoja, according to several studies, are effective only in the short term and very
destructive in the long run (Stark, 2011; Stites and Akabwai, 2009).
In Karamoja, the economy and land have developed under constant outside
pressure. Evolution and adaptation to these changes have forced many Karamojong to
migrate to nearby cities. Others have diversified into charcoal and livestock trade. Tree
cutting and burning for commercial charcoal fuel has gained prominence in the area.
Deforestation is making the region more vulnerable and will likely lead to further civil
unrest, land wrangles and other stressors, which in turn lead to further environmental
degradation, conflict and underdevelopment (Morton, 2001).
Today, population growth and harsh climate conditions are leading to greater
competition for food, water, and pasture. Information on alternative integrated
approaches to managing resulting food insecurity is becoming more crucial to secure
lives. Efforts to induce development have been unsuccessful in the past due to the
failure to take into account the complexity of the people’s culture and society. Instead,
interventions have exacerbated problems leading to conflict and undermining of the
resilience of pastoralism. There is urgent need to understand the true needs of the
society. Local knowledge should be utilized to formulate sustainable approaches
7
capable of overcoming the true hurdles faced by the people of Karamoja.
Across East Africa, droughts have shaped and continue to play a key role in
shaping the human and natural environment of the region (Juma, 2009). This is true
in Karamoja where droughts are considered the driver and modulator of conflict
(Stark, 2011), food insecurity (Nakalembe, 2017), land use and land use policies
(Nakalembe et al., 2017). A significant amount of research often qualitative has been
conducted documenting drought and its impacts on Karamoja. However, very few
studies have quantitatively assessed drought vulnerability in Uganda or even
characterized drought in Karamoja, the region considered to be most prone to
droughts. The research conducted as part of this dissertation seeks to quantitatively
examine drought, quantify agricultural land use expansion from satellite data,
document the impact of programs and policies promoting crop-based livelihoods and to
document local perceptions of drought through interviews with local farmers in a
region that is considered extremely vulnerable to drought.
1.2 Research Objectives
The overarching goal of my research is to better understand vulnerability to
agricultural drought and to quantitatively assess the impacts of drought on the people
of Karamoja of northeastern Uganda. In support of this primary goal the issues
addressed are summarized in Figure 1.1. The specific objectives of the research were:
1. To quantitatively assess vulnerability to drought in Uganda using a multiscale
and spatially explicit method;
8
Food Aid Land use policy (settlement) 	
Livelihoods 
Chronic Food Insecurity 
Land use 
Sedentary crop based production	
Vulnerability 
(Exposure +  Sensitivity)/ Adaptive Capacity 
Karamojong Community Perceptions	
Drought	
Figure 1.1: Schematic of issues addressed in this thesis
2. To quantify agricultural expansion from satellite data and document the
dominant drivers leading to the current state of agricultural land use;
3. To characterize agricultural drought in Karamoja to enable a comprehensive
understanding of drought and its impacts, concomitantly evaluating the
suitability of satellite-based drought monitoring;
4. Document farmers’ perceptions of drought in the region.
1.3 Dissertation Organization
This dissertation is organized into four major sections each corresponding to the
specific objectives listed in Section 1.2. Chapters 2 through 5 have been written in
9
such a way that they can stand alone for publication. These Chapters correspond to
the objectives above and for this reason, some information contained in the
introduction and study area sections may be repetitive. Chapter 2 assesses
sub-regional vulnerability to drought in Uganda using a spatially-explicit and scalable
method. Chapter 3 examines cropland expansion within Karamoja in Uganda, the
most vulnerable to drought in the country. This was done by investigating the links
between biophysical and political/historical events leading to the current state of
agricultural land use. This chapter was published in the Land Use Policy Journal in
March 2017 (Nakalembe et al., 2017). Chapters 2 and 3 set the stage for a detailed
assessment of agricultural drought in Karamoja, Uganda. Chapter 4 characterizes
drought and its impact in Karamoja, Uganda and proposes a method to estimate the
number of people in need of food aid based on drought severity. This chapter is
currently under review by the Natural Hazards Journal (Nakalembe, 2017). Chapter 5
focusing only on Moroto District in Karamoja, uses qualitative research methods to
better understand the human perceptions of drought. Chapter 6 includes a brief
summary of the dissertation research and results, areas for future research and the
policy implications of the findings.
10
2 Mapping drought vulnerability in
Uganda
2.1 Summary
The research in this chapter presents a multiscale, spatially explicit, quantitative
approach to drought vulnerability assessment within Uganda. The three dimensions of
vulnerability: exposure, sensitivity, and adaptive capacity; are analyzed for Uganda’s
11 sub-regions using rainfall (station and satellite derived), remote-sensing and
socio-economic data from the most recent Uganda National Panel Survey (UNPS) of
2014 1. The methods presented in this chapter are scalable and easily reproducible,
highlighting opportunities for strategic interventions. Although the Karamoja Region
is moderately exposed to drought, it is highly sensitive and by far has the least
adaptive capacity in Uganda. The region is, therefore, the most vulnerable to
agricultural drought. Adaptive capacity is the main contributor to variation in drought
1UNPS data are available on the World Bank Living Standards Measurement Study (LSMS) Website:
http://go.worldbank.org/FS2M7AYE00
11
vulnerability between regions. The results indicate that unless strong programs are put
in place to increase the adaptive capacity (particularly, poverty alleviation in Northern
Uganda), severe drought impacts will continue to devastate the region. This study also
highlights the need for a more detailed assessment of drought at the sub-regional level
in Karamoja to better understand its temporal and spatial severity across the region
where more and more agro-pastoralists are reverting to pure crop-based livelihoods
aided by policy and programming.
2.2 Introduction
“Global change vulnerability is the likelihood that a specific coupled
human-environment system will experience harm from exposure to stresses
associated with alterations of societies and the environment, accounting for
the process of adaptation” (Schro¨ter et al., 2004)
The impact of drought on any society depends on its vulnerability, and while
droughts may be seasonal and have the same severity, intensity and spatial extent, the
impacts on society vary from drought to drought and from region to region (Smakhtin
and Schipper, 2008; Wilhite, 1985). Vulnerability is a measure of how susceptible a
society is to a hazard and its associated effects (Howe, 2009; Smakhtin and Schipper,
2008). Vulnerability is context specific and varies by community. For instance, low
adaptive capacity (e.g. lack of income diversity, entrenched poverty and dependence on
rain-fed agriculture) makes most of Sub-Saharan Africa vulnerable to drought. The
projected increase in the frequency of extreme climate events in the region is likely to
12
increase its vulnerability (Egeru et al., 2014; Barrios et al., 2008; McCarthy et al.,
2001). Vulnerability to drought is also increasing in many regions in the world due to
mounting pressure on water and other natural resources particularly in marginal areas
(Wilhite et al., 2007).
In response to increasing severity and impacts of hazards, disaster management
often focuses on response and recovery, with little or no attention to mitigation,
preparedness, prediction and/or monitoring. There is need to shift to risk management
approaches that necessitate analyzing vulnerability (Fontaine and Steinemann, 2009;
Wilhelmi and Wilhite, 2002b). In Uganda, the National Policy for Disaster
Preparedness and Management formulated in 2010 clearly indicated a departure from
the previous approaches. The policy among other objectives aims to establish and
equip disaster preparedness and management institutions at national and local
government levels to generate and disseminate information on early warning for
disasters and hazard trend analysis (Directorate of Relief Disaster Preparedness and
Refugees, 2010). Irrespective of the policy, persistent losses from disasters in Uganda,
in particular from drought and flooding, suggest growing vulnerability. This also points
to inadequate resources and planning to fast track realization of the policy’s objectives.
Analyzing vulnerability enables identification of appropriate measures to mitigate
potential damage when hazards occur (Wilhelmi and Wilhite, 2002b). Since
vulnerability is multifaceted; many frameworks of vulnerability assessment exist today.
These include the use of causality by modeling potential exposure to hazard events,
focusing on underlying social conditions, and those that integrate both (Cutter et al.,
2008). Assessments often vary depending on the region, scale and data availability
13
(Cutter et al., 2008; Adger, 2006). For example, Wilhelmi and Wilhite (2002b)
investigating vulnerability to agricultural drought in Nebraska in the United States by
assessing two biophysical factors, climate and soils, and two other factors, land use and
irrigation Prelog and Miller (2013) evaluated vulnerability based on three social and
ecological elements; the presence of hazards, individual-level characteristics, and
community-level variables. While studies including; McNeeley et al. (2016); Porter
et al. (2014); Murthy et al. (2014); Liu et al. (2013); Cinner et al. (2012); Antwi-Agyei
et al. (2012); Howe (2009); Fontaine and Steinemann (2009) and Allison et al. (2009)
characterize vulnerability as a function of the exposure, sensitivity and adaptive
capacity.
In Allison et al. (2009), Uganda’s fisheries sector is ranked 16 of 33 of the most
highly vulnerable to projected climate change under the Intergovernmental Panel on
Climate Change (IPCC) scenario B22. In Brooks et al. (2005), Uganda ranked in the
moderate to highly vulnerable to climate variability. Although this information is
critical at the global level and when comparing countries, it offers very little
information on vulnerability variability across regions in Uganda. At the sub-national
level vulnerability studies including Helgeson et al. (2013); Ide et al. (2014); Okonya
et al. (2013) and Hisali et al. (2011) have focused primarily on coping strategies and
adaptive capacity of different regions and have not thoroughly explored the exposure
and sensitivity components of vulnerability. While the above point to a growing body
of case studies on vulnerability assessments in Uganda, they also point to the need for
2Scenario B2 describes a world in which the emphasis is on local solutions to economic, social, and
environmental sustainability (IPCC, 2000)
14
quantitative assessments that are scalable and easily reproducible using readily
available data. This is in addition to the pressing need for comprehensive data on
impacts and occurrences of hazards.
Assessing both vulnerability and human perception of drought can provide a
more comprehensive understanding of communities when designing resilience programs
to reduce potential impacts (van Aalst et al., 2008; Cutter et al., 2008; Adger, 2006).
The objective of this study was to use readily available data to assess vulnerability in
Uganda that would enable comparability between regions and in time to assess any
changes in vulnerability over time. Of particular interest was to assess how regions
responded to drought conditions e.g. Karamoja fit in the national context.
2.3 Study area
Uganda, located in East Africa has a total land area of 200.523 km2 and a total
population of 34.6 million people with an annual averaged growth rate of 3% (Uganda
Bureau of Statistics, 2016) (Figure 2.1). Uganda’s population is growing at an
unprecedented rate and it is one of the main drivers of land use and environmental
change. At 34,634,650 in 2014, Uganda’s population increased by an estimated 10
million people since 2002 (Uganda Bureau of Statistics, 2016). At this rate, the
population of Uganda is projected to reach 55 million by 2030 and 100 million by 2050.
Figure 2.2 shows the population trends in Uganda from 1911 to 2014 (Uganda Bureau
of Statistics, 2016). Population pressure is increasing land fragmentation, which will
inevitably lead to a reduction in per capita production, increased income inequality
15
Figure 2.1: Study Area: Uganda
and inevitably increased vulnerability to drought (UNDP, 2007).
Uganda is a landlocked country; therefore, the Inter-Tropical Convergence Zone
(ITCZ) and air currents (i.e., the southeast and northeast monsoons) influence its
equatorial climate (National Environment Management Authority, 2008). The central,
western and eastern regions have two marked rainfall seasons; March to May (MAM)
and September to November (SON). In these regions average annual rainfall exceeds
2000mm (Figure 2.3). The northern region however, receives one rainy season from
April to October with the north east (Karamoja) receiving the least amount of rainfall
averaging 700mm per year. Vegetation varies from tropical rain forest in the southwest
to savannah woodlands and semi-arid vegetation in the northeast following the rainfall
gradient. The livelihood zones are defined by major crops and land use practices.
16
1911 1921 1931 1948 1959 1969 1980 1991 2002 2014 2020* 2030* 2050*
0
10
20
30
40
50
60
70
80
90
100
110
2.5 2.9 3.5
5
6.4
9.5
12.6
16.7
24.2
34.6
41.4
55.6
100.4
Census year
P
op
u
la
ti
o
n
(’
00
0
,0
00
)
Figure 2.2: Population trends in Uganda from 1911 to 2014.*=projected population
These are similarly influenced by rainfall distribution. For example, the drier areas
UG05, UG21 grow sorghum that is more suited for dry conditions, while the wetter
central regions UG26 and UG39 grow bananas (Figure 2.4).
Uganda’s arable land is estimated to be over 70% and agriculture is aided by
fertile soils and regular rainfall for most of the country. Despite the sector’s
importance, the use of agricultural inputs such as improved seeds, fertilizer, and
pesticides is still limited (Uganda Bureau of Statistics, 2013a).
Uganda faces a wide range of development challenges including low life
expectancy, extreme poverty, and high levels of income inequality (Uganda Bureau of
Statistics, 2013a). Although the Human Development Index (HDI)3 has been
3The Human Development Index (HDI), a composite statistic of life expectancy, education, and per
capita income indicators, which are used to rank countries into four tiers of human development
17
Entebbe Gulu
Kabale Kotido
Arua Soroti
Figure 2.3: Average monthly rainfall at selected stations.
Figure 2.4: Uganda livelihood zones (FEWS NET, 2015)
18
increasing, Uganda remains in the low human development category, ranked 163rd of
188 countries in 2014 (UNDP, 2015). Poverty4 was estimated at 20% and most of
Uganda is rural with a 79% of households (over 26 million people) living in rural areas
(Uganda Bureau of Statistics, 2016).
Approximately 80% of households in Uganda are involved in agriculture (Uganda
Bureau of Statistics, 2016). Despite a meager 20% contribution to the national gross
domestic product (GDP), agriculture remains the most important sector with 64.7% of
the country’s working population (aged 14 to 64) engaged in subsistence rain-fed
agriculture (National Environment Management Authority, 2008; Uganda Bureau of
Statistics, 2016). Average total agricultural land cultivated is very low, with the
national average at 1.1 Ha, the highest being in the Northern region at 1.6 and the
lowest in the Western region at 0.8 Ha (Uganda Bureau of Statistics, 2010a).
Uganda’s mean annual temperature has been increasing at an average rate of
0.28 ◦C per decade since 1960. It is projected to increase by 1 ◦C to 3 ◦C by the 2060s
with an increasing frequency of hot days especially during June, July, August and
September (McSweeney et al., 2015). This increase, particularly in the northeast where
air temperatures are high and a 1.5 ◦C increase is projected, will likely amplify water
shortages and the impact of droughts on agriculture (see for example Figure 2.5) (Funk
et al., 2012). Annual rainfall is projected to increase during October, November
December by up to 35% and a reduction during June, July and August with an
http://hdr.undp.org/en/content/human-development-index-hdi
4Absolute Poverty Line is equivalent to one US dollar per person per day in purchasing power parity
expressed in 2005/06 prices
19
(a) (b) (c)
Figure 2.5: A farmer stands in his sorghum field that was destroyed by the 2015 drought
in Rupa, sub-country, Moroto District Karamoja (a). Destroyed maize (b)
and beans (c) in Lwengo and Rakai District respectively and were destroyed
by the 2016 drought.
increase in extreme events (rainfall that falls in heavy events) during the MAM and
OND rainfall seasons (Christensen et al., 2007; McSweeney et al., 2015). However,
model projections disagree on the projected changes of future El Nin˜o and La Nina
events that significantly affect rainfall in East Africa (Anyamba and Tucker, 2002;
McSweeney et al., 2015).
Drought impacts are not well documented in Uganda, which limits understanding
of societal vulnerability and inevitable, and the ability of officials to respond
adequately to drought events or to allocate resources in advance of an event (Wilhite
et al., 2007). It is important to note that, vulnerability is dynamic and changes based
on the economic, social, and environmental characteristic of a region. As with other
natural hazards, drought mitigation and preparedness are key to the reduction of
future impacts (Wilhite et al., 2007). This necessitates a timely, continuous and
20
quantitative assessment of vulnerability.
2.4 Data and methods
2.4.1 Assessing drought vulnerability in Uganda
Research in this chapter applies a multi-scale, spatially-explicit, quantitative
approach to drought vulnerability assessment within Uganda at the sub-regional level.
It combines rainfall data (station and satellite-derived), remote-sensing information
(NDVI) and socio-economic data from the most recent Uganda National Panel Survey
(UNPS) of 2014. The data used in this study are summarized in Table 2.1.
Vulnerability has three dimensions: Exposure, Sensitivity and Adaptive capacity
(Fontaine and Steinemann, 2009; Schro¨ter et al., 2004) (Equation 1);
V = (E + S)/AC (2.1)
where E = exposure to drought (severity and frequency), S = sensitivity of
cropped areas to agricultural drought (drought impacts in agriculture), and AC =
adaptive capacity of regions to cope with drought (determined using socioeconomic
proxy indicators from the UNPS data).
2.4.2 Assessing exposure
In this study, exposure is measured in terms of the severity and frequency of
droughts at the sub-regional scale using the Standardized Precipitation Index (SPI)
21
Data Timescale Source Index calculated Measure
Climate Station Data 1960 to 2012 UNMA Monthly rainfall
CHIRPS 2012 to 2016 CHG Monthly rainfall -
Station and CHRIPS Data 1960 to 2016 UNMA & CHG SPI Exposure
MODIS NDVI 2001 to 2016 GLAM East Africa NDVICV, NDVIMax Sensitivity
UNPS 2014 LSMS Poverty rates, land ownership, etc Adaptive Capacity
MODIS NDVI 2012 to 2016 GLAM East Africa Moroto NDVI Anomaly Experience
Semistructured interviews 2014 - - Definitions, expections
Table 2.1: Data use to assess exposure, sensitivity and adaptive capacity
(McKee et al., 1993). SPI values were calculated from rainfall data from twelve climate
stations across the country obtained from the Uganda National Meteorological
Authority (UNMA) (1960 to 20012). Rainfall data from 2013 to 2016 were predicted
from a simple regression using station and the Climate Hazards Group InfraRed
Precipitation (CHIRPS) data (1983 to 2016) (Funk et al., 2015). SPI is calculated as
the difference in precipitation from the mean of a specified time period divided by the
historical mean and standard deviation allowing analysts to measure the rarity of a
drought event at a particular time scale (e.g. 1, 2, 3, 6, 9, 12 or 18 months) (McKee
et al., 1993). In this study, we used the 3-month time scale that has been applied in
measuring drought impacts in agriculture (Ji and Peters, 2003). Exposure in this study
accounts for frequency and severity of agricultural drought. The severity or drought
categories are based on the SPI-thresholds adopted from the United States Drought
Monitor5 summarized in Table 2.2 (McKee et al., 1993). Therefore, each region’s
exposure index is calculated as;
5More information at: http://droughtmonitor.unl.edu/aboutus/classificationscheme.aspx
22
4∑
i=1
f ∗Di
100
(2.2)
where Di= drought severity (Table 1), f = frequency
2.4.3 Assessing sensitivity
Sensitivity to drought is the degree of susceptibility of agricultural areas within
the region to moisture stress (Murthy et al., 2014). In this study, sensitivity to drought
is calculated from sub-regional cumulative and maximum values of satellite-based
Normalized Difference Vegetation Index (NDVI) data. Remote-sensing data have been
used extensively in studies assessing drought since the early 1980’s. Recent drought
studies include (Gouveia et al., 2016; Vicente-Serrano et al., 2015; Zhang and Jia,
2013b; Mishra et al., 2015) and (Ji and Peters, 2003). These data have also been used
widely in studies of flooding (Sanyal and Lu, 2005), landslides (Metternicht et al.,
2005; van Westen et al., 2008), soil erosion (Rahman et al., 2009) while multi-hazards
studies include (Huggel et al., 2002; Tralli et al., 2005; Alcantara-Ayala, 2002). These
data provide timely, high quality and spatially explicit data for such assessments
especially for countries that lack databases on impacts of hazards.
In this study, 250 meter MODIS NDVI data were obtained from the Global
Agriculture Monitoring System (GLAM) for East Africa from 2001 to 2016
(Becker-Reshef et al., 2010a). The MODIS NDVI data were sub-set using sub-regional
polygons from Uganda National Bureau of Statistics (UBOS) 2014 to obtain values
pertaining to each sub-region from the larger MODIS NDVI image. Analysis
23
Category Description Possible Impacts SPI-Value
D0 Abnormally dry Going into drought: short-term dryness -0.5 to -0.7
slowing planting, growth of crops or pastures
some lingering water deficits
pastures or crops not fully recovered
D1 Moderate Some damage to crops, pastures -0.8 to -1.2
Streams, reservoirs, or wells low
some water shortages developing or imminent
D2 Severe Crop or pasture losses likely -1.3 to -1.5
Water shortages common
Water restrictions imposed
D3 Extreme Major crop/pasture losses -1.6 to -1.9
Widespread water shortages or restrictions
D4 Exceptional Exceptional and widespread crop/pasture losses -2.0 or less
Shortages of water in reservoirs,
streams, and wells creating water emergencies
Table 2.2: Drought categories and possible impacts adapted from United States Drought
Monitor
24
concentrated at the peak of the growing season MAM and SON for bi-modal regions
and JJA for Karamoja.
2.4.4 Assessing adaptive capacity to cope with drought
Drought adaptive capacity is the ability of individuals to reduce the adverse
effects of drought through actions taken before, during or after drought (McCarthy
et al., 2001). The IPCC 2001 report asserts that determinants of adaptive capacity are
highly correlated with measures of economic development. Therefore, this study
selected indicators from the most recent Uganda National Panel Survey (UNPS) of
2013-2014 including indicators on land ownership, sources of income and size of plots
(Adger, 2006; McCarthy et al., 2001). The UNPS is carried out annually on a
nationally-representative sample of households (Uganda Bureau of Statistics, 2014).
The survey is multi-purpose and allows exploration of the determinants of adaptive
capacity. Descriptive statistics were analyzed for each of the variables by region. Each
region was ranked 1 to 10 (1 = least adaptive capacity and 10= highest adaptive
capacity). We assigned larger weights (higher adaptive capacity) to non-drought
sensitive sources of income including wage employment, non-agricultural enterprises,
and property income. Lower weights (lower adaptive capacity) were assigned to
subsistence farming and commercial farming which are more drought-sensitive.
25
2.5 Results
2.5.1 Exposure to drought in Uganda
Sub-regional sensitivities to drought in Uganda are summarized in Figure 2.6.
The results indicate that the broader northern region is by far the most exposed to
meteorological drought. Teso is the only sub-region that falls in the extremely exposed
class followed by West Nile and Lango, both highly exposed. Central 2 is by far the
least exposed. In terms of totals, the Karamoja region experienced the most droughts.
However, the majority (twelve in total) were in the mild (abnormally dry) category
between 1960 and 2016. Lango experienced the highest number of exceptional droughts
(twelve in total) (Table 2.3). The impacts of exceptional droughts include widespread
crop/pasture losses, shortage of water in reservoirs, streams, and wells often causing
water emergencies (see Table 2.2) (United States Drought Monitor, 2015).
2.5.2 Sensitivity to agricultural drought in Uganda
Karamoja and Teso region are by far the most sensitive to drought in Uganda
(Figure 2.7). This implies that in both regions, meteorological droughts often lead to
severe impacts on crops (agricultural drought) depicted by the very low cumulative
growing season and maximum NDVI values. It is no surprise that sorghum, which is
drought-tolerant is grown in both regions (FEWS NET, 2015). Elgon, Lango and
Western region are also highly sensitive to agricultural drought, followed by Acholi,
Central 1, and Eastern which are moderately sensitive. Central 2 and South Western
have low sensitivity to agricultural drought, while West Nile is the least sensitive.
26
Figure 2.6: Sub-regional exposure to drought in Uganda
From these results, it appears that in Karamoja even mild droughts (droughts in the
abnormally dry, moderately dry categories) often have severe impacts on agriculture.
2.5.3 Sub-regional adaptive capacity to cope with drought impacts
in Uganda
Sub-regional poverty rates
Adaptive capacity was assessed by analyzing descriptive statistics of selected
indicators from the UNPS. The data on poverty rates indicates that Karamoja at 66%
has the highest poverty rates. It is followed by Eastern, which includes Teso and Elgon
and the mid-north, which include Acholi and Lango. Central 1 (excludes Kampala),
27
Sub-region (Station location) D0 D1 D2 D3 D4 Exposure
Acholi (Gulu Station) 7 4 10 2 3 0.68
Central 1 (Kampala Station) 5 6 4 0 8 0.69
Central 2 (Jinja Station) 10 5 2 0 6 0.56
East Central (Tororo) 6 5 6 1 6 0.68
Elgon (Tororo) 6 5 6 1 6 0.68
Karamoja (Kotido Station) 12 3 7 1 5 0.68
Lango (Lira Station) 6 4 0 0 12 0.74
South Western (Mbarara & Kabale) 7 2 4 1 7 0.66
Teso (Soroti) 4 5 8 1 7 0.77
West Nile (Arua Station) 8 5 3 0 9 0.72
Western (Kasese & Masindi Stations) 7 3 2 1 8 0.63
Table 2.3: Sub-regional exposure to drought in Uganda (where D0-Mild, D1-Moderate,
D2-Severe, D3-Extreme, and D4-Exceptional Drought)
Central 2 followed by South Western are the least poor (Table 2.4).
Sub-regional parcel sizes and land ownership
Shown in Table 2.5 is a summary of land ownership by sub-region. It can be seen
from this table that on average land parcel sizes are largest in Central 1 although the
majority 41% only own one acre. East Central had the highest number of households
(6.8%) with more than 6 acres of land.
28
Figure 2.7: Sub-regional sensitivity to drought in Uganda
sub-region Non-poor Poor Total Sample Percent poor Rank
Acholi Mid-North 252 133 385 34.5 4
Central 1 342 29 371 7.8 11
Central 2 290 27 317 8.5 10
East Central 228 92 320 28.8 6
Elgon (Eastern) 258 167 425 39.3 2
Karamoja (North East) 35 68 103 66 1
Lango Mid-North 252 133 385 34.6 5
South-westrn 379 49 428 11.5 8
Teso (Eastern) 258 167 425 39.3 2
West Nile 204 70 274 25.6 8
Western (Mid-West) 257 29 286 10.1 7
Table 2.4: Sub-regional poverty rates
29
Region No land I acre 2 to 5 Acres 6+ Acres Rank
Acholi Mid-North 20.69 37.07 39.66 2.59 6
Central 1 26.11 41.40 27.39 5.10 11
Central 2 21.33 37.33 38.00 3.33 8
East Central 25.00 40.91 27.27 6.82 4
Elgon (Eastern) 38.95 41.58 17.37 2.11 2
Karamoja (North East) 25.00 45.00 25.00 5.00 1
Lango (Mid-North) 20.69 37.07 39.66 2.59 6
South-westrn 31.58 39.18 28.07 1.17 5
Teso (Eastern) 38.95 41.58 17.37 2.11 2
West Nile 40.14 37.32 19.01 3.52 9
Western (Mid-West) 20.69 37.07 39.66 2.59 10
Table 2.5: Land ownership by sub-region shown as percent of sample from the Uganda
National Panel Survey of 2014
30
Figure 2.8: Sub-regional adaptive capacity to drought in Uganda based on social eco-
nomic indicators from the Uganda National Panel Survey of 2014
Overall adaptive capacity
Overall adaptive capacity by sub-region in Uganda is shown in Figure 2.8.
Karamoja has the lowest adaptive capacity with a total score of 0.17. The region had
the highest rates of poverty, lowest education levels, and the very high dependency on
rain-fed agriculture (Table 2.6 and Table 2.7). For all but 2 indicators (income from
wages and income sources), Karamoja ranked at the bottom. Lango and Acholi follow
in the low category with 0.38 and 0.39 respectively. While Central 1 followed by
Central 2 had the highest adaptive scores of 1.02 and 0.89 respectively.
31
Table 2.6: Summary of indicators of adaptive capacity used in this study from the
Uganda National Panel Survey of 2014
Indicator Acholi Central1 Central2 East Central Elgon
Source of earnings 1 11 9 7 5
Income from wage, salary, etc 1 10 11 8 4
Soil/land quality= poor 2 10 11 5 8
Main water source to parcel = Irrigation 5 11 8 10 2
Inadequate household stocks of food due to drought 7 9 2 6 4
Formal certificate of title or customary ownership 7 10 11 9 5
Highest grade/class completed 3 9 8 7 10
Parcel acquisition= purchased 2 10 11 7 8
Land ownership 6 11 8 4 2
Poverty rates 4 11 10 6 2
Overall adaptive capacity 0.38 1.02 0.89 0.69 0.5
Table 2.7: Summary of indicators of adaptive capacity used in this study from the
Uganda National Panel Survey of 2014
Indicator Karamoja Lango South Western Teso West Nile Western
Source of earnings 3 1 4 5 h 10 8
Income from wage, salary, etc 6 1 7 4 3 9
Soil/land quality= poor 1 2 7 8 4 6
Main water source to parcel = Irrigation 1 5 4 2 7 9
Inadequate household stocks of food due to drought 1 7 2 4 11 10
Formal certificate of title or customary ownership 1 7 4 5 2 3
Highest grade/class completed 1 3 5 10 6 2
Parcel acquisition= purchased 1 2 4 8 6 5
Land ownership 1 6 5 2 9 10
Poverty rates 1 5 8 2 8 7
Overall adaptive capacity 0.17 0.39 0.5 0.5 0.66 0.69
32
Figure 2.9: Sub-regional vulnerability to agricultural drought in Uganda
2.5.4 Vulnerability
Overall vulnerability of sub-regions in Uganda is shown Figure 2.9. The broader
northern Uganda region (which includes; Acholi, Elgon, Karamoja, Teso and West
Nile) is by far the most vulnerable to drought in the country. This vulnerability is
mostly driven by the very low adaptive capacity. Karamoja sub-region is the most
vulnerable followed by Acholi, Lango, Teso, and Elgon in the high vulnerability
category. West Nile, western, south western fall in the moderate vulnerability category,
while Central 1 and Central 2 fall in the very low category due to high adaptive
capacity; the two are also the most urbanized and developed. Individual scores by
region are summarized in Figure 2.10.
33
Exposure Sensitivity Adaptive-capacity Vulnerability/4
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
1.1
1.2
S
c
o
r
e
Acholi Central-1 Central-2 East Central Elgon Karamoja
Lango South-Western Teso West-Nile Western
Figure 2.10: Summary of exposure, sensitivity, adaptive capacity, vulnerability to drought in Uganda
34
2.6 Conclusions
This study developed and applied an easily reproducible and scalable
quantitative method to analyze the relative vulnerability to drought in Uganda using
readily- available data.
Drought exposure does not vary extensively across Uganda, however, sensitivity
and adaptive capacity vary greatly. It is clear that adaptive capacity more specifically
socioeconomic development contributes the most to this vulnerability. The general lack
of livelihood alternatives in the most drought sensitive regions coupled with poor
adaptive capacity makes Karamoja 2.5 times more vulnerable to drought than any
other sub-region in Uganda.
The results strongly indicate that unless strong programs are put in place to
increase adaptive capacity, particularly focusing on eradicating poverty in northern
Uganda through alternative livelihood streams, severe drought impacts in Uganda will
continue to devastate the region. The Government of Uganda recognizes that northern
Uganda is lagging behind due to protracted conflict and violent cattle rustling. During
the 1990’s the Lord’s Resistance Army (LRA) rebel group often raided and destroyed
villages, affecting most of Acholi, West Nile, and Lango and during 2010-2011,
Karamoja experienced violent cattle rustling.
The Government of Uganda has been implementing programs to accelerate
development in the region. These include: the Peace Recovery and Development Plan
(PRDP), Northern Uganda Social Action Fund (NUSAF) and Karamoja Livelihoods
Programme (KALIP) initiatives. The programs appear to fall short of their goals given
35
the continued severe drought impacts.
There is a need for increased attention toward improving drought management at
national and sub-national level. A key first step would include the compilation of
comprehensive and reliable information on drought impacts. This information would
be essential for improved understanding of Uganda’s vulnerability to drought. It would
also justify the need for increased investment in less drought sensitive sectors or devise
mitigation strategies. Without this information, it will remain difficult to convince
policy and other decision-makers of the need for additional investments in drought
monitoring and prediction, mitigation, and preparedness.
36
3 Agricultural Land Use Change in
Karamoja Region, Uganda
3.1 Summary
We examine dramatic cropland expansion in Karamoja, Uganda by investigating
the links between biophysical and political historical events leading to the current state
of agricultural land use. Our objective was to quantify agricultural expansion, uncover
the dominant drivers leading to the current state of agricultural land-use and its
impacts on livelihoods. Region-wide analysis of remotely-sensed data, land-use policy
and history as well as farmer interviews were undertaken. We posit that government
programs instituting sedentary agriculture are the most significant drivers of cropland
expansion in Karamoja. We show a 299% increase in cropland area between 2000 and
2011 with most expansion occurring in Moroto District (from 706 hectares to 23,328
hectares). We found no evidence of an increase in overall crop production or food
security and food aid continues to be essential due to recurrent crop failures. Due to
lack of resources for inputs (e.g., seeds and labour) cultivated fields remain very small
37
in size and over 55% of once cultivated land is left fallow. Our findings bring into
question whether continued promotion of rain-fed agriculture in Karamoja serves the
best interests of the people. Current cropland expansion is directly competing and
compromising pasture areas critical for livestock-based livelihoods. Without strong
agricultural extension programs and major investments in climate-smart options,
cropland expansion will continue to have a net negative impact, especially in the
context of current climate projections which indicate a future decrease in rainfall,
increase in temperature and an increase in the frequency and magnitude of extreme
events.
3.2 Introduction
Since the 1960’s, underlying forces (e.g., cropland expansion programs, controlled
grazing) originating from land use policy and development programs, other than
proximate causes (direct local level actions), have profoundly influenced land-cover and
land-use change in Karamoja. Policies based on modernization theory developed
during colonial and post-independence periods have greatly influenced what Karamoja
is today. Modernization theory advocated for privatization and individualism in the
region, which systematically discouraged the traditional customary1 mutual right to
grazing land, whilst promoting agricultural expansion and commercial mining
(Mwebaza, 2015). Since the 1960’s customary land was increasingly devoted to state
1Customary tenure recognizes both individual, family and communal ownership of land (Coldham,
2000)
38
use for mining, forestry, wildlife conservation and mineral exploration (Rugadya and
Kamusiime, 2013). These policies contradicted and directly suppressed pastoralism,
the traditional livelihood in the region.
Introduction of regional borders in 1984 disorganized pastoralists by limiting
their movement and access to formally alternative communal grazing lands during
drought years (Stites and Akabwai, 2009; Gelsdorf et al., 2012). This escalated cattle
raids (Gray, 2000), land grabbing (Kra¨tli, 2010) and overgrazing (Mamdani, 1982).
Today, the majority of the people of Karamoja have very little control over land
ownership, its use and/or management is augmented by government policy that limit
communities’ rights over the land (Rugadya and Kamusiime, 2013; Gelsdorf et al.,
2012). Population, policy, biophysical and economic factors have and continue to
impact land cover and land use in Karamoja resulting into complex consequences on
livelihoods and the environment. Examining this complexity requires a deeper look
into what underpins land use and change in the region.
Today, two ideologies about suitable livelihoods in Karamoja clash. While
literature advocates for pastoral livestock-based livelihoods e.g., Levine (2010); Avery
(2014); Sundal (2010); Matthysen and Finardi (2010), government programs are
promoting sedentary crop-based livelihoods often leading to cropland expansion with
the aim of increasing production in spite of the erratic climate and continued crop
failures. Levine (2010) indicates that the best livelihood strategy for most of Karamoja
is livestock-based herding, yet majority of government and development programs give
more support to crop-based livelihoods.
To examine factors leading to the current state of agricultural land-use and
39
expansion in Karamoja, we begin our analysis with a brief background of the region,
demography and livelihoods. We then detail land use history and policy highlighting
circumstances that underpin instituted agricultural development and the current state
of agricultural land use in the region. We quantitatively assess agricultural/cropland
expansion in the region using satellite data and present a new high-resolution cropland
map for Karamoja for 2011/12. Cropland-based agricultural land use is investigated
specifically because there have been extensive programs aiming at increasing crop
production primarily through area expansion but to our best knowledge no real
quantifiable positive outcomes such as increased production or food security have been
realized. Our analysis gives equivalent importance to the cultural, economic, political
and biophysical aspects to enable a robust understanding of instituted policy impacts
on agricultural land use and change in Karamoja.
3.3 Study Area: The Karamoja Region of Northeastern
Uganda
Located in northeastern Uganda, Karamoja encompasses 28,000 square
kilometers between 1-4o North and 33-35o East (Figure 3.1) with the total population
estimated at 1,017,900 million people2. It is projected that 74.5% (about 700,000
people) live below the national absolute poverty line compared to only 19.7% for the
2UBOS, 2014 Census Provisional Results
40
Figure 3.1: Karamoja Livelihood Zones 2015 Source: Integrated Food Security Phase
Classification (IPC) Report
41
rest of the country (Uganda Bureau of Statistics, 2013b)3.
Karamoja is set on a large plateau at an average elevation of approximately 1,000
meters above sea level. The land plain rises to northeast towards the hilly terrain
bordering the escarpment above the neighboring Turkana District in Kenya. The
underlying basement complex rock mostly consists of undifferentiated acid and
granitoid gneisses. Four preeminent mountains intersperse the plains, Mt. Morungole
in the north, Mt. Kadam in the south, Napak the largest mountain in the region is in
the southwest, and Mt. Moroto, in the east. The Karamoja plains are dominated by
sandy clay loams of low agricultural production. Some parts are degraded by soil
erosion with low water holding capacity, while the base of Napak and Kadam
Mountains is characterized by volcanic soils of medium productivity (Luzinda and
Wilson, 1959; Kagan et al., 2009).
Rainfall in Karamoja is highly variable and sporadic in space and time ranging
from 350mm to 1,500mm per year (Office of the Prime Minister, 2010). This broad
range means both dry events (drought) and wet events (flooding) can occur in close
proximity. This results in a characteristic heterogeneity of weather, vegetation and
crop performance. The high environmental variability (vegetation, soils and terrain)
across the region coupled with a dependence on rainfall significantly impacts crop
yields in Karamoja. In general, crop yields are highly variable in space and between
years, and are currently impossible to predict. Landscape location (riverbed vs.
foothills), wealth status and farmer practices also impact crop yields. Our fieldwork
3The official absolute poverty line (which is equivalent to one US dollar per person per day in Pur-
chasing Power Parity (PPP)) expressed in 2005/06 prices
42
revealed that within one village in the same season, yields of sorghum could range from
zero (complete failure) to one ton per acre (exceptional).
The rainfall gradient determines major livelihood and land use characteristics.
Defined by the dominant combination of land-use, Karamoja is classified into three
main livelihood zones excluding game reserves, Kidepo National Park and urban areas
as shown in Figure 3.1 along a gradient of decreasing rainfall from the west towards
the drier and more agriculturally marginal northeast. As shown in Figure 1, the
livelihood zones are simply an indication of the most significant contributor towards
livelihood sustenance and are not mutually exclusive. The agricultural livelihood zone
(K02) is dominated by crop cultivation with minimal dependency on livestock for food
and income. Field sizes are very small, estimated at 1.6 hectares4 and are often mixed
cropped combinations of sorghum, maize, beans and/or millet. Sorghum and maize are
by far the dominant crops and the most important for food security. The agro-pastoral
zone (K03 and K04) is drier than the agricultural zone but receives enough rainfall to
sustain crop cultivation complemented by livestock rearing. To manage uncertainty by
ensuring access to water and pasture during the dry season, agro-pastoralists used to
move between semi-permanent villages called ”manyattas” and mobile cattle camps
called ”kraals” (Stites and Akabwai, 2009; Niamir-Fuller, 1999). The pastoral zone
(K05) is the driest of the three zones and livestock and livestock products’ sales are the
dominant sources of income.
To the naked eye, Karamoja looks virtually unchanged since the 1960’s and the
landscape remains sprinkled with grass-thatched manyattas (Figure 3.2). However, at
4UBOS 2010. Uganda census of agriculture 2008/2009, Volume IV: Crop Area and Production Report
43
closer examination, population growth, and external forces including government
programs, and demand for mineral resources have left significant impacts. Some of the
changes including agricultural expansion, a decrease in woody vegetation, an increase
in surface mining of gold and marble and greater clustering of manyattas are
manifestations of the combined effects of development programs, insecurity, population
pressure and culture. To-date, understanding of agricultural land use in Karamoja
remains largely anecdotal and to our knowledge there have been no empirical studies of
agricultural land use, its expansion and adoption by this formerly largely pastoral
society. Understanding the characteristics of these changes, including when and where,
the shortcomings and the underlying motivations for current practices can provide
insight into the trajectory of changes necessary for future planning for sustainable
development within and beyond the agricultural sector in marginal Karamoja.
3.4 Land use history
According to Rugadya and Kamusiime (2013) Karamoja was home to large herds
of wildlife including buffalos, elands, zebras, topis, hartebeests, giraffes and elephants
that enabled hunting during the 1920s. By the 1950s, wildlife populations had declined
tremendously, which compelled the government to gazette most of the land as
protected areas. By 1962, 94.6% of land in Karamoja (mostly communal and pastoral
land) had been allocated to wildlife conservation, forestry, and mineral exploration and
prospecting (Rugadya and Kamusiime, 2013). Kidepo National Park and three
44
(a) (b)
(c)
Figure 3.2: (a) A 1960 Arial Photo of A Manyatta in Karamoja (Deshler, 1960). Mod-
ern day Karamoja, An aerial photomosaic of a Manyatta of Pupu village,
Pupu parish, Rupa sub-county (b) [Photo taken August 20, 2015] using an
unmanned aerial vehicle (UAV) with permission from Chief Executive Offi-
cer, Moroto District and the Uganda Peoples Defense Forces (UPDF). Kids
stand in front of their home in Kadilakeny, Rupa (c) [Photo taken January
21, 2016]
45
controlled hunting areas in Napak, North Karamoja and South Karamoja were
established in 1963, and in 1964 three additional game reserves in Matheniko, Bokora
Corridor and Pian-Upe were designated (Rugadya and Kamusiime, 2013).
Earlier in 1894 when Uganda became a British Protectorate, the newly-formed
arbitrary administrative districts impacted pastoral communities (Sundal, 2010). For
example according to (Sundal, 2010) 15% of Karamoja’s land was lost to the
Protectorate of Kenya as a result. This disorganized pastoral communities through
grazing zone and movement restrictions that limited cross-border grazing into
neighboring Districts and into Kenya. In 1921, when Karamoja officially became a
province with a civil administration, seasonal livestock migration was restricted to
dry-season grazing lands (Stites and Akabwai, 2009; Gelsdorf et al., 2012; Gray, 2000;
Sundal, 2010). Limited movement within Karamoja meant limited access to grazing
land in the drought prone region, escalating cattle raids, land grabbing, and depletion
of natural resources due to excessive exploitation (Matthysen and Finardi, 2010;
Gelsdorf et al., 2012). The net impacts of colonial programs particularly Karamoja
Cattle Trading scheme were; the reduction in cattle herds and raised income for the
government. On the other hand the livelihood for the people of Karamoja was
diminished due to the decrease in herd sizes, increase of violent raids, increase of taxes,
decrease of grazing areas and increase of cattle disease epidemics (Johnson, Douglas
Hamilton Anderson, 1988; Mamdani, 1982; Sundal, 2010).
Insecurity due to cattle raids has been associated with pastoralism in East
Africa, initially arising from worsening pasture conditions leading to pasture searches
beyond traditional areas (Mamdani, 1982). Insecurity escalated with intensified
46
Figure 3.3: Land use history 1960 to 2002: Conservation areas dropped from 94.6% in
1965 to 40.8% in 2002 (Rugadya and Kamusiime, 2013). Degazetting further
negatively impacted pastoralists by curbing access to pasture in the newly
designated wildlife reserves but opened up more land to sedentary agriculture
(Kra¨tli, 2010)
47
cross-border raids due to increased access to illegal firearms within the region after the
Ugandan army soldiers abandoned their barracks with an estimated 60,000 weapons in
1979 (Mirzeler and Young, 2000; Ayoo et al., 2013; Stark, 2011). By the 1990’s cattle
herds were built through raids that became extremely violent (Mugerwa, 1992)5.
Insecurity, limited access to pasture and water sources exacerbated land
degradation, and increased cattle losses from raids, escalating poverty levels (Kington,
2002; Gray et al., 2003; Gray, 2000; Bevan, 2008). Mirzeler and Young (2000)
indicated that following the 1980 drought and associated famine, the livestock
population in Karamoja had dropped by more than half to 300,000 from the 600,000
estimate of 1962. As a result of violent raids and drought in 1990 50 to 60% of
households did not own livestock and many households had shifted to crop cultivation.
Ethnic groups including Pokot, Chekwe, Labwor and Dodoth were reported to have
abandoned cattle due to raids but the Jie, Matheniko, Bokora and Pian still depended
on the pastoral economy (Mugerwa, 1992; Burns et al., 2013).
In 2002 due to heavy encroachment by pastoralists into protected areas, the
Government of Uganda (GoU) approved to de-gazette 14,904 square kilometers (60%)
of previously gazetted land, to allow for expansion of crop farming and commercial
mining (Rugadya and Kamusiime, 2013; Kra¨tli, 2010). Degazetted areas included part
of Pian-Upe, Bokora Corridor and Methenico wildlife reserves that were reduced while,
community wildlife areas were established in Iriri, Karenga and Amudat (Rugadya and
5For a detailed history on traditional cattle raiding that turned commercial to pay taxes and extremely
violent upon increase in small arms in the region see Matthysen and Finardi (2010); Bevan (2008);
Mirzeler and Young (2000); Stites et al. (2007)
48
Kamusiime, 2013). In summary, conservation areas dropped from 94.6% in 1965 to
40.8% in 2002 (Figure 3.3) (Rugadya and Kamusiime, 2013). Degazetting further
negatively impacted pastoralists who lost access to pasture areas that were designated
wildlife reserves but opened up more land to sedentary agriculture (Kra¨tli, 2010;
Rugadya and Kamusiime, 2013).
Today, even with improved security more homesteads are still being built closer
to each (Stites et al., 2007). This has created a unique and distinct pattern on the
landscape as village gardens surround manyattas with well-established definitive
boundaries marked by trees, dry-bushes and/or rocks (Figure 2). Communities engage
in mining, firewood collection and charcoal production to cope with low crop yields
and the loss of animal assets. The products are sold or bartered for food or local brew
residue at town markets. Wild fruits and vegetables are also collected for family
consumption (Stites and Akabwai, 2009).
Development program agendas have also significantly contributed to and
influenced the rates of agricultural land use change. One of the earliest programs was
the Karamoja Resettlement/Rehabilitation Scheme of 1955-1958, designed to develop
Karamoja preceding Uganda’s independence by allocating resources to develop new
water supplies including wells, ponds and valley dams and introduce progressive
methods of land use such as farming and establishment of meatpacking plants (Sundal,
2010). The Karamoja Integrated Disarmament and Development Program (KIDDP)
initiated in 2004, led to the disbandment of traditional kraals replacing them with
protected kraals established alongside army barracks for safekeeping of livestock.
Protected kraals restricted the movement of livestock causing depletion of pasture and
49
water resources and resulted in animal losses to diseases (OCHA, 2009; Stites et al.,
2010; Burns et al., 2013). Also under the KIDDP, a tractor-hire scheme distributed an
estimated 29 heavy-duty tractors and seven walking tractors in Karamoja and
according to Nsibambi (2014) an estimated 8,396 acres of land were ploughed and
planted through the scheme and distributed 950 metric tons of planting materials to
local farmers in Karamoja. This scheme that started in 2011 was suspended in 2013
due to poor performance, mismanagement and for having no impact on yields
(Kasasira, 2013).
Other programs including The Karamoja Action Plan for Food Security
(KAPFS), The Peace, Recovery and Development Plan for Northern Uganda (PRDP)
(Government of Uganda, 2007; Office of the Prime Minister, 2009) and Northern
Uganda Social Action Fund Project (NUSAF) now in its third phase. Though these
programs have not been independently evaluated, there is very little if any evidence to
suggest they achieved their objectives in Karamoja. For example, KAPFS
implemented in 2009 was intended to ensure sustainable food security and to increase
household income yet most of Karamoja is far from it. It is clear however that KAPFS
prioritized crop over livestock production. This is exhibited in the KAPFS finance plan
for intervention budget which allocated 43.9 million Uganda shillings to objective 1
which aimed at increasing crop production compared to only 8 million towards
objective 2, aimed at increasing livestock production and productivity (Office of the
Prime Minister, 2009).
Data about human population dynamics in Karamoja are scarce and often
unreliable, however available data indicate per capita livestock numbers have not
50
Figure 3.4: Population dynamics in Karamoja. Source data: MAAIF (2010); Mirzeler
and Young (2000); UBOS (2004, 2009); Uganda Bureau of Statistics (2010b,
2002); Burns et al. (2013)
significantly changed since 1950, though the number of households engaged in
agriculture has been increasing in the region. The Ministry of Agriculture, Animal
Industry and Fisheries (MAAIF) reported in 2010 that food-crop acreage continued to
expand as a direct result of the disarmament program in Karamoja. Households
engaging in crop cultivation have increased in-spite of recurring droughts and
unreliable rainfall that cause crop failures and systematic famines consequently
increasing the dependency on food aid (Figure 3.4). Figure 4 shows that the estimated
population needing food aid has remained above 50%, livestock per capita ownership
has remained lower than the 1959 estimate, but population as well as the area under
cultivation has been steadily increasing.
The sequence of events summarized above offer a view into the evolution of land
51
use towards more crop-based livelihood in Karamoja resulting from a complex
interaction of policy interventions and culture. In the next sections we summarize
findings from interviews held with heads of farmer households in Moroto District to
establish the trends in farming practices.
3.5 Agricultural land use in Moroto District Karamoja
To gain insight into the trends of farming practices within a sample of farmers in
Moroto district, a purposive sample of twenty farmer household heads including 14
men and 6 women from 14 villages in Rupa, Nadunget and Katikekile sub-counties
were interviewed. Interviewees were asked standard questions to assess their farming
experience: their age, the duration they have been farming, the main crops they grow,
land holdings, average yield, and amount sold to the market.
Of the 20 farmers interviewed, 14 (74%) had been farmers for more than 15
years, 3 (16%) for 5 to 15 years while only 2 (11%) had been farmers for less than 5
years. All the farmers interviewed stated that they mainly grow sorghum, maize and
beans. 29% also grew sunflower, and 38% grew vegetables - mostly Sukuma Wiki (a
kale-like green). Other crops grown include groundnuts, simsim, cowpeas, green grams,
pumpkins and tomatoes. When asked what they considered a good harvest, sorghum
averaged 1.8 mt/ha. Maize averaged 1.2 mt/ha, while beans averaged 0.5 mt/ha. 8
farmers (about 40%) stipulated that a good harvest is when they are able to fill 3 to 4
granaries which on averages accommodates 0.3 to 0.5 mt of sorghum6.
6Personal communication, Mr. Olinga John, Moroto District Agricultural Officer, December 5, 2015
52
(a) (b)
Figure 3.5: An aerial image of a perfectly demarcated 0.394 hectares field in Rupa sub-
county, Moroto District (a) and (b) is a photos of sorghum fields in Moroto
District including a field ploughed under the tractor hire scheme (February
2011) (top left)
In comparison to the broader Northern Uganda and to the national production as
reported in the 2008/2009 census of Agriculture 2008/2009 crop production in
Karamoja as a whole is very poor (Table 3.1). Maize production for Karamoja
averaged 0.70 mt/ha below the 1.2 mt/ha for Northern Uganda and the national
average of 2.3mt/ha(Uganda Bureau of Statistics, 2010b). Karamoja sorghum yields
stood at only 0.38 mt/ha compared to Northern Uganda at 0.7 and the national at 0.9
mt/ha however more consistently and spatially disaggregated data is much needed to
provide a representative picture of production in the region.
The interviews also revealed that present day cultivation within the sampled
group is without inputs such as fertilizer and there is no irrigation, despite the fact
that all farmers indicated unreliable rainfall as their number one challenge. Other
53
Area (ha) Production (mt) Yield (mt/ha) % of National Production
Maize
Moroto District (included Napak) 3,755.00 3,736.00 0.83 0.2
Karamoja 26,177.00 18,432.00 0.70 0.78
Northern Uganda 247,780.00 305,798.00 1.23 12.9
National 1,014,260.00 2,361,956.00 2.33 100
Survey average 0.40 0.48 1.19
Sorghum
Moroto District (Included Napak) 14,290.00 11,332.00 0.79 3.0
Karamoja 117,406.00 44,333.00 0.38 1.88
Northern Uganda 249,330.00 177,088.00 0.71 47.1
National 399,252.00 375,795.00 0.94 100
Survey average 0.40 0.76 1.88
Table 3.1: Moroto District production data compared to Northern Uganda (which in-
cludes Karamoja region) and national production as reported in the 2008/9
census of Agriculture (Uganda Bureau of Statistics, 2010b)
54
stated challenges included pest and diseases and only one farmer indicated insecurity
due to cattle rustling as a major challenge. Agriculture within the study population is
still primarily for subsistence. This was shown by the fact that 11 (55%) of the
interviewed households indicated that they do not sell any of their produce, 7 (35%)
stated that they sell between 25 to 50% of their produce and only 2 (10%) replied that
they sell more than 50% of their produce. Average field size was 2 acres (0.809371
hectares), which is consistent with national statistics reported by UBOS. Figure 3.5 is
an aerial image of a perfectly demarcated 0.394 hectares field in Rupa sub-county. For
the most part, farmers in in the sample group engage in other activities than farming
to sustain their livelihoods. 14 (70%) of the farmers we interviewed indicated that they
owned livestock and 3 (15%) indicated that they brew and sell local beer, while only 1
(5%) stated that they engage in casual labor, such as stone quarrying and mining.
Only 10% indicated that they engage in cutting trees for firewood and charcoal
production, especially during drought years.
While the sample was too small and data are too few to draw statistically
representative conclusions, they revealed that crop cultivation is not new in Moroto.
The majority 14 (74%) had been engaged in some form of crop cultivation since
childhood. In context with the land-use history of the area, crop cultivation was of
secondary importance to livelihood sustenance as many owned and depended on
livestock. The data confirm our observation that crop cultivation in Moroto remains
for subsistence and productivity remains very low.
The section on land-use history in combination with the data presented above
gives an insight into agricultural land use in Karamoja. It also hints at the landscape
55
scale changes that have likely happened however, does not offer any explanations of the
landscape scale impacts. In the next section we use satellite data to quantify cropland
area expansion that is inevitably leading to loss of pastureland and follow with
synthesized explanations of the current state.
3.6 Estimating cropland expansion in Karamoja with
remote sensing data
Advances in remote sensing technology and methods during the course of the last
30 years, have seen development of new data products and tools necessary to map and
monitor even the smallest and most remote agriculture. Product and data quality have
improved primarily due to improved pre-processing algorithms, the increase in sensors
for land monitoring and increased spatial resolution (Pflugmacher et al., 2011). Table
3.2 summarizes some of the land cover products that have been generated for Africa at
the continental and global scale from satellite data which include the region of study.
Fritz et al. (2015) offers detailed descriptions of the methods and attributes of some of
the products listed below. It is important to note that there is great uncertainty in the
cropland distribution over Africa in global to continental scale products and their use
has to be with extreme caution (Fritz et al., 2010).
Accessing very high-resolution satellite data needed to map and derive land cover
information in highly heterogeneous landscapes with smallholder agriculture is still a
challenge. As shown in Figure 3.6, data that are freely available including Landsat and
56
Dataset(Sensor) Date (Resolution) Source Reference
IIASA/IFPRI Cropland 2005 1 km IIASA (Various) (Fritz et al., 2015)
Forest Cover Change 2013 (30 m) Global Forest Watch (Hansen et al., 2013)
GLOBECOVER 2009 2010 (300 m) ESA (http://due.esrin.esa.int/globcover/)
GLOBECOVER 2006 2006 (300 m) ESA (http://due.esrin.esa.int/globcover/)
MODIS Land Cover 2001-2012 (0.5ox0.5o) GLCF/UMD (http://glcf.umd.edu/data/lc/)
GLC-2000 (SPOT-4) 2000 (1 km) JRC,EC (Murad and Islam, 2011)
MOD12V1 (MODIS Terra) 2000-2001 (1 km) Boston University (Friedl et al., 2002)
AFRICOVER (Landsat) 2000 (30 m) FAO (Gregorio and Jansen, 1998)
Uganda Land Cover Scheme II 2000 & 2014 (30m) RCMRDSERVIR http://apps.rcmrd.org/landcoverviewer/
Table 3.2: Examples of land cover products available for Africa, for details see Fritz
et al. (2010)
MODIS are too coarse and are limited when deriving land cover information in
Karamoja. This is due to the very small size of fields (averaging 1.3 hectares), as well
as fuzzy field boundaries and the similarity of fallow fields with surrounding
uncultivated grasslands. In addition the lack of good regional scale land cover and land
use information further constrains estimating biophysical changes. Hence, to examine
the changes in spatial extent of croplands in Karamoja we derived a
2010/2011-cropland map from WorldView 1&2 panchromatic and multi-spectral data
at 0.5 m and 1.8 m resolution respectively. The satellite images were acquired between
2010 and 2011 and obtained from The National Geospatial-Intelligence Agency (NGA)
through an agreement with NASA under the NextView License7. Figure 3.7 shows the
footprints of 185 images with the least-percent cloud cover selected from over 1136
7For information about the NextView License, please visit: http://cad4nasa.gsfc.nasa.gov/
57
Figure 3.6: Spatial details obtainable from different satellite systems. With WorldView
2 data at 2m spatial resolution, individual fields (averaging 1.6 hectares)
can be delineated. This is not feasible at lower spatial resolution i.e. using
Landsat (30m) and MODIS (250m) data.
images for digitization.
The radiometrically corrected Worldview data were converted to
top-of-atmosphere spectral reflectance, and orthorectified using the DLR/SRTM
Digital Elevation Model (DEM) to correct for relief and terrain effects prior to
digitization. The data were organized into a GIS database for easy retrieval, analysis
and visual interpretation. On-screen digitization was undertaken on each of the 185
selected images using visual interpretation of agriculture versus non-agriculture 3.8.
Post classification comparison was done to estimate change in cropland area between
2000 and 2011/12 by comparing the Uganda Land Cover Scheme II classification with
the newly created data product. A post-classification (map to map) comparison was
performed to generate a difference map and a matrix of change (Singh, 1989) and to
identify large-scale changes that might be explained by ancillary data. Field visits were
carried out in 2011, 2013, 2014 and 2015 for initial reconnaissance and to collect
training and validation data.
58
Figure 3.7: Footprints of 185 Worldview 1 and Worldview 2 Images with the least-
percent cloud cover selected from over 1136 images for digitization. Data
Source-Digital Globe: NextView License
59
Figure 3.8: Derived 2012 croplands map showing the extent of agricultural land use in
Karamoja
60
3.7 Cropland expansion in Karamoja
Year (Spatial Resolution) Uganda Land Cover Scheme II 2000 (30m) New Croplands 2011 /12 ( 2˜m) Percent Change
Abim 10,715.70 32,271.30 201
Amudat 4,494.10 2,293.50 -49
Kaabong 29,966.60 68,911.30 130
Kotido 20,014.10 86,974.70 335
Moroto 706.1 23,328.70 3204
Nakapiripirit 7,915.40 66,272.30 737
Napak 14,838.70 81,756.60 451
Total (Ha) 90,650.70 361,808.40 299
Table 3.3: Karamoja cropland area in hectares and percent change estimates between
2000 and 2011/12
Cropland area and percent change by District is shown in Table 3.3. Of a total
2,800,000 hectares only 90,650.70 (3.24%) was cropland in 2000 Figure (3.9A), which
expanded to 361,808.40 (12.92%) by 2011/12 Figure 9(B), representing a 300%
increase in cropland area between 2000 and 2011/12 Figure 9(C). Most agricultural
expansion occurred near major towns including Nabilatuk in Nakapiripirit, Kaabong
town in Kaabong and Kangole in Napak that can be seen in the derived
post-classification difference map Figure 8(C). We show that extensive expansion in
croplands has occurred in all other Districts in Karamoja with the exception of
Amudat District where cropland area reduced by 2,200.6 hectares. We anticipate this
might be an error of commission in the 2000 data, which necessitates further field
investigation and accuracy assessment.
61
(a) (b) (c)
Figure 3.9: Cropland Area 2000 (SERVIR USAID) (a), New Cropland Area Map 2011/12 (b) and New Croplands area high-
lighted in (c)
62
District Major Crops Area 2008/9 Mapped Croplands 2010/12 Percent Cultivated 2008/9
Abim 3,825.00 86,974.00 4.40
Kaabong 28,015.00 105,084.00 26.66
Kotido 27,526.00 68,565.80 40.15
Moroto (Includes Napak) 18,345.00 32,271.30 56.85
Nakapiripirit (Includes Amudat) 69,171.00 68,911.30 100.38
Table 3.4: Estimate of percent area under cultivation in Karamoja
When we compare the 2011/12 map data to official crop area and production
report of 2010 from the 2008/9 census of agriculture we can assume that there was a
55% increase in cultivated land in the 3 year period or that up to 55% of once
cultivated land was left fallow based on reported total area of major crops (Maize,
finger millet and sorghum). With the exception of Nakapiripirit (at the time included
Amudat) other district were less than 50% cultivated (Table 3.4). Integrating these
findings with our survey and average household area planted data from UBOS, we
deduce that households lack the resources (labour and inputs) to cultivate more than
the average 2 acres and also rotate plots frequently. Future work will necessitate using
spatially disaggregated production data and mapping pasture areas that have been
most likely converted to croplands.
3.8 Summary of factors contributing to cropland
expansion in Karamoja
Government programs are the primary driver of cropland expansion in Karamoja.
Much of the expansion in the region can be explained by program driven increased
63
importance of crop production over pastoralism. Programs and policies led to
reductions in grazing areas and decrease in livestock numbers whilst advocating and
investing in crop cultivation. More land has hence been converted and used for crop
cultivation.
Another factor for rapid expansion of cropland areas in between 2000 and 2012 in
Karamoja is increased security due to disarmament that started in 2006. This lead to
the opening of new areas for cultivation and explains patterns of croplands distribution
in Karamoja whereby, a majority of croplands are located close to and/or surround
protected kraals and Manyattas. As a result natural resources close to settled areas are
already over-stretched or depleted leaving many with no other alternative but farming.
As a consequence of loss of livestock to diseases and sales, many households have
no other alternative but to engage in crop cultivation. And although overall the total
number of livestock has been increasing (Figure 4), per capita ownership has remained
low. Today fewer households own and depend on livestock and for many crop
cultivation remains the only alternative. According to a recent food security and
nutrition assessment by WFP (Figure (3.10)) only 8% of households in Karamoja
owned greater than 5 Tropical Livestock Units (TLU)8 and except for Amudat district
at 8% over 40% of households in other districts did not own any livestock (World Food
Program, 2016).
The impact of population growth on cropland expansion cannot be understated.
Irrespective of civil unrest and extreme hardship, Karamoja’s population has been
growing, almost doubling from 370,000 in 1991 to 699,000 in 2004, and increasing yet
81 TLU is equivalent to 1 250 kg tropical cow
64
Figure 3.10: 2016 livestock ownership in Karamoja (World Food Program, 2016)
again to 1,017,000 in 2008. Access to cultivable land is still very high and is estimated
at 1.2 hectares per household as compared to 1.1 nationally (Mwesigwa et al., 2014;
Uganda Bureau of Statistics, 2010b). In Karamoja today, an individual can simply
start cultivating an open area9 and as the population grows, demand for food will
increase. This will continue to drive many to grow crops and conversion of this easily
accessible land.
3.9 Conclusions
Our analysis of satellite data show extensive croplands expansion between 2000
and 2011/12. The rate of this change is so dramatic and we conclude has been driven
largely by agricultural development programs. These programs also provided inputs
9Personal communication Mr. Olinga John (District Agricultural Officer and Mr. Edonu Janan
(District Entomologist)- Moroto District May 12, 2016
65
such as seeds that are distributed at the beginning of the growing season and put new
land under the plow. Hence, expansion was unequivocal and we anticipate that this
trend will continue, pending continued investment by the government and development
agencies.
Uganda has no official policy on land use or agro-pastoralism, but through
development programming (e.g Karamoja Cattle Trading Scheme, KAPFS, PRDP and
KIDDP) the relative significance of crop cultivation for sustaining livelihoods has
increased for many households in Karamoja while dependency on livestock has been
reduced. Both the current livelihood zone map and our land use map show that areas
formerly mapped as purely pastoral currently grow crops. However, due to the nature
of soils with very low water holding capacity, insignificant investments in irrigation
(Avery, 2014) and poor access to inputs crop production makes meager contributions
to livelihood sustenance.
Although cultivated area has dramatically increased, we find no quantifiable
overall increase in yield, or per-capita production due to the recurrent poor food
security. This status quo, (poor yields and dependence on food aid) is likely to be
maintained as more land is put to crop cultivation by poor households and meager
investments are made in livestock based livelihood opportunities. From our study it
appears that water for irrigation has to be secured for crop cultivation to impact on
food security given the climate in the region and the nature of soils. This is likely
unattainable due to the high costs and low water availability (Avery, 2014). Although
small scale irrigation schemes are littered (Figure 3.11) across the region their
contribution is very minimal and is yet to be quantified (Avery, 2014).
66
Figure 3.11: Vegetable and maize gardens under drip irrigation in Nadiket, Moroto
January 2016
Cropland areas have expanded but due to the low capacity and lack of inputs
necessary to efficiently use the land, a majority of the land is left fallow and percent
land utilization is very low. In recent years food aid has become paramount for
sustenance of life in the region where pastoralism is hindered and the climate is too
erratic to support sustainable rainfed food crop cultivation. This hints at the likely
loss and destruction of large pasture areas have been cleared for croplands and not
utilized resulting in a no-win situation for both pastoral and agro-based households.
Cropland expansion is also reinforcing deforestation, competing with grazing
concomitantly compromising ecosystem services in marginal Karamoja.
The process, patterns and impacts of land use change remain of secondary
importance and if current programs are continued, agricultural expansion will
continue, although the future rates of change and extent are unknown. Due to the lack
of market incentives, crop production and cropland expansion driven by development
programs are likely to continue. However, overall production is likely to remain poor.
Finding a balance between crop-based (where suitable) with livestock-based livelihood
67
should be of primary importance in development planning for Karamoja and will
require empirical understanding of economic, environmental and ideological differences
which should guide development planning and investments.
68
4 Characterizing Agricultural Drought
in the Karamoja sub-region of
Uganda with Meteorological and
Satellite-Based Indices
4.1 Summary
Karamoja is notoriously food-insecure and has been in need of food aid for most
years during the last two decades. One of the main factors causing food insecurity is
drought. Reliable, area-wide, long-term data for detecting and monitoring drought
conditions are critical for timely, life-saving interventions and the long-term
development of the region, yet such data are sparse or unavailable. Due to advances in
satellite remote sensing, characterizing drought in data sparse regions like Karamoja
has become possible. In this study, we characterize agricultural drought in Karamoja
69
to enable a comprehensive understanding of drought, concomitantly evaluating the
suitability of NDVI-based drought monitoring. We found that in comparison to the
existing data, NDVI data currently provide the best, consistent and spatially explicit
information for operational drought monitoring in Karamoja. Results indicate that the
most extreme agricultural drought in recent years occurred in 2009 followed by 2004
and 2002 and suggest that in Karamoja, moderate to severe droughts (e.g. 2008) often
have the same impact on crops and human needs (e.g. food aid) as extreme droughts
(e.g. 2009). We present in a proof-of-concept frame, a method to estimate the number
of people needing food aid and the population likely to fall under the Integrated Food
Security Phase Classification (IPC) Phase 3 (crisis) due to drought severity. Our
model indicates that 90.7% of the variation in the number of people needing aid can be
explained by NDVI data and NDVI data can augment these estimates. We conclude
that the biggest drivers of food insecurity are the cultivation of crops on marginal land
with insignificant inputs, the lack of irrigation and previous systematic incapacitation
of livestock (pastoral) alternatives through government programing. Further research is
needed to bridge empirical results with social economic studies on drought impacts on
communities in the region to better understand additional factors that will need to be
addressed to ensure livelihood resilience.
4.2 Introduction
Current socio-economic conditions in pastoral (agro-pastoral) East Africa bring
into question the worthiness of aid programs and policies in marginalized and
70
poorly-understood regions. Pastoralists in Africa are still considered the continent’s
most vulnerable group. To this day despite continued inflow of aid and poverty
alleviation programs, extreme vulnerability due to entrenched poverty is vividly
manifested during recurrent droughts that continue to have severe detrimental impacts
on communities (Little et al., 2008; Juma, 2009).
Although droughts are often labeled as the root cause of food insecurity in the
Karamoja sub-region of North Eastern Uganda, the evidence is inconclusive
(Nakalembe et al., 2017; Ayoo et al., 2013; Stark, 2011; Ferreri et al., 2011).
Agriculture in the region is primarily for subsistence, largely characterized by very
small fields with minimal to no agricultural inputs such as fertilizer. Through
systematic government programs, crop cultivation has taken precedence over livestock
keeping, yet the contribution of crop cultivation to food security remains meager
(Nakalembe et al., 2017). Droughts stress livelihoods in Karamoja, often with
devastating consequences however, the reasons for near stagnant if not declining
resilience to drought in the region are complex. It is evident that the effects of
droughts such as famine, escalating conflict, displacement, migration and the loss of
livestock and human lives remain a big concern in Karamoja (Gelsdorf et al., 2012;
FEWSNET, 2005).
The characteristics and impacts of drought vary within Karamoja (Egeru et al.,
2014). On the ground it is evident that small amounts of rainfall in one village, which
lead to complete crop failure, can coincide with sufficient amounts for a decent harvest
in another village within a few kilometers’ distance (Figure 4.1). Prior to the
completion of the government’s disarmament campaigns that started in 2001, drought
71
(a) (b)
Figure 4.1: (a). A failed sorghum field in Rupa Sub-county, 13 August 2015. (b) A
moderately impacted field in Nadunget Sub-county, 14 August 2016. These
two fields are within 10 kilometers distance.
induced food insecurity and its impacts were also escalated by conflict and vice-versa
in the region (FEWSNET, 2005). Irrespective of the long history of drought being
blamed for famines and social unrest, there is currently no monitoring system in place.
Uganda has no nationally run drought early warning and monitoring system;
currently there is no basis to measure the relative severity of individual drought
occurrences. Furthermore, early drought detection, tracking, mapping and severity
assessment are considerably constrained by the lack of meteorological data and very
few empirical studies quantitatively assess drought at the sub-regional level,
particularly assessing drought and it’s impacts of agriculture. This explains the lack of
consensus on the characteristics of drought in Karamoja today. With the exception of
a few empirical studies such as Egeru et al. (2014), the vast majority including
Gelsdorf et al. (2012), Stark (2011), DanChurchAid (2010), and Stites et al. (2007)),
focus primarily on the social and economic impacts of drought with minimal empirical
backing, but nevertheless, remain the only source of information for policy-making.
72
FEWSNET (2005) for example, reported that droughts that used to occur every five
years now occur every two to three years. Based on Normalized Difference Vegetation
Index (NDVI) analysis FEWSNET (2005) concluded that severe droughts were
experienced in 1992, 1994, 1998, 2000, 2001, 2002, and 2004. On the other hand, Stites
et al. (2007) notes that droughts occur every three to four years in Karamoja, with
multi-year droughts every 10 years, while Ayoo et al. (2013) indicate that Karamoja
experiences cyclical droughts every two to three years. Using such information can
mislead decision-making and planning.
There are several drought monitoring systems developed by development
partners including the Famine Early Warning Systems Network (FEWS NET) created
by US Agency for International Development (USAID) that provides early warning
and analysis on food insecurity and World Food Program’s (WFP) Vulnerability
Assessment and Mapping (VAM) Seasonal Explorer (World Food Programme, 2017).
FEWS NET’s provided objective and evidence based information from combing
satellite data and expertise from the National Aeronautics and Space Administration
(NASA), National Oceanic and Atmospheric Administration (NOAA), United States
Department of Agriculture (USDA), and United States Geological Survey (USGS) with
field information from regional scientist and to date provides some of the best
information about food insecurity and drought in East Africa (FEWSNET, 2017).
Both websites are a rich resource, particularly the FEWS NET website that provides
tools, data, assessments and reports but to the best of authors knowledge, these are
not largely used by government departments in Uganda for their reporting and/or
monitoring.
73
There is also a general lack of empirical studies of drought and it’s impacts on
agriculture in Karamoja. This remains an obstacle to agricultural development in
Karamoja. This information is now more critical with the current and projected rates
of extreme events (McSweeney et al., 2015). The goal of this study is to demonstrate
the opportunity for a nationally run operational drought monitoring for Karamoja, by
using readily available remote-sensing information to derive spatially-explicit drought
information in combination with historical climate station data to empirically
characterize agricultural drought at the sub-regional level. This information is essential
in the design and implementation of emergency response programs as well as food
security and Social Safety Net programs.
4.2.1 Defining and characterizing drought
Drought remains one of the most complex and devastating, but also least
understood natural disaster (Wilhelmi and Wilhite, 2002a). Drought intensity,
variability, duration and impacts differ from drought to drought and from region to
region, hence no single universal drought definition exists (Edwards, 1997). Drought
definitions should identify regional differences in terms of specific climatic conditions,
physiological characteristics, economic development and traditions to understand the
potential impacts on livelihoods (Kogan, 1995). However, in many resource and data
scarce regions droughts are poorly defined, which calls into question the suitability of
drought management programs (Alley, 1984; Wilhite, 1985; Heim, 2002;
Vicente-Serrano et al., 2010).
74
According to Palmer (1965), a meteorological drought is an interval (generally
spanning several months or years) during which the actual moisture supply at a given
place consistently falls below the expected or climatically adequate moisture supply.
However, the climatically adequate moisture supply is difficult to determine and
average rainfall alone does not provide an adequate statistical measure to determine
drought, particularly in drier areas (Palmer, 1965; AghaKouchak et al., 2015).
Moreover, the same precipitation deficit could have different impacts depending on
other meteorological factors, vegetation types, farming practice and the sociopolitical
situation of the affected area. Therefore, droughts are grouped into four main types
and are often defined according to disciplinary perspectives (Heim, 2002; Wilhite,
1985).
In addition to meteorological drought defined above, drought types include
agricultural, hydrological, and socioeconomic. Moisture deficits (meteorological
drought) that occur at critical growth stages for crops leading to dryness in the root
zone can lead to agricultural drought that severely reduces crop yields (Heim, 2002;
Palmer, 1965). A hydrological drought occurs when precipitation deficits affect surface
or subsurface water supply, thus reducing stream flow, groundwater, reservoir, and lake
levels, while socioeconomic drought associates the supply and demand of some
economic good with elements of meteorological, agricultural, and hydrological
droughts. Hence, though difficult, drought characteristics and impacts are not
impossible to identify and quantify plus, recent advances in land remote-sensing make
possible spatially explicit drought studies even in the most data scarce regions
(Wilhite, 1985).
75
To-date several meteorological drought indices have been used to systematically
study droughts. These include the Palmer Drought Severity Index (PDSI) (Palmer,
1965), Precipitation Anomaly Percentage (PAP) (Yang and Wu, 2010) and the
Standardized Precipitation Index (SPI) (McKee et al., 1993), based on the cumulative
probability of a given rainfall event at a given meteorological station (Vicente-Serrano
et al., 2010; National Drought Mitigation Center, 2015). These indices are often
location-specific and although they provide accurate point-based estimates of drought
conditions, they are limited when climate data are lacking or incomplete and limited in
providing spatially-continuous drought information across large areas (Quiring and
Ganesh, 2010; Brown and DeBeurs, 2008). Moreover, in semi-arid regions the spatial
heterogeneity of rainfall events makes the use of point-based rainfall estimates to
calculate these indices highly questionable since these data are extrapolated over large
distances between rain gauge stations (Flitcroft et al., 1989).
The Normalized Difference Vegetation Index (NDVI) remains the most
commonly used satellite-based index used to monitor greenness of land surfaces, plant
vigor, density and growth conditions (Atzberger, 2013; Quiring and Ganesh, 2010;
Karnieli et al., 2010; Tucker, 1979). The Vegetation Condition Index (VCI) (Kogan,
1995) and Standardized Vegetation Index (SVI) both derivatives of NDVI and the
Normalized Difference Water Index (NDWI) (Gao, 1996) are the most commonly used
in drought studies. VCI is per-pixel normalized NDVI to enhance the weather-related
component in NDVI time-series data by reducing influence of environmental conditions
and has been used to characterize drought at various spatial and temporal scales (Jain
et al., 2009; Ji and Peters, 2003; Anyamba et al., 2001; Kogan, 1995). The SVI is the
76
probability of vegetation conditions diversion from the normal (Peters et al., 2002).
These satellite-based indices have an advantage over station-based indicators since
spatial details on vegetation condition are inherent (Kogan, 1995).
Combinations of various multi-source, multi-sensor and multi-index information
capitalizing on the advantages of individual indices to conduct multipurpose,
multi-temporal and multi-spatial scale studies tend to produce more reliable drought
information (Xu et al., 2011; Suwanprasert et al., 2013). Several studies have
investigated the relationship between satellite and meteorologically-based indicators of
drought to produce more accurate information. Liu and Kogan (1996) showed that
drought patterns delineated by NDVI and VCI agree well with rainfall anomalies
observed from rainfall maps, providing useful information to analyze quantitatively the
temporal and spatial characteristics of regional drought and crop production.
Moreover, strong correlations between satellite indices and crop yield have been
demonstrated, particularly during the critical periods of crop growth (Franch et al.,
2015; Lotsch et al., 2003; Kogan, 1997). Franch et al. (2015) proved that timeliness of
reliable forecast of winter wheat production prior to the NDVI peak can be improved
when growing degree information from climate data is added to the Becker-Reshef
et al. (2010a) method that combined NDVI, official reported crop statistics and crop
type masks. Lotsch et al. (2003) demonstrated a high degree of association between
NDVI and SPI time series. Kogan (1997) also showed that VCI could be used to
identify vegetation stress even in the early growing season and proved to be useful for
real-time assessments, diagnosis of vegetation condition and weather impacts on
vegetation. In addition, Peters et al. (2002) indicated that SVI when used with other
77
indicators is a good indicator of short-term weather conditions and can be used as an
indicator of onset, extent, intensity and duration of stress.
Several factors including rainfall regime, vegetation and soil types affect the time
response of NDVI to available rainfall amount (Quiring and Ganesh, 2010; Liu and
Kogan, 1996). Rainfall and moisture availability in particular influence the lagged
relationship between NDVI and meteorological indices. For example, Quiring and
Ganesh (2010) show that in Texas the growing-season VCI responds to prolonged
moisture stress and is less sensitive to short-term precipitation deficiencies. Ji and
Peters (2003) similarly demonstrated the lagged relationship between weather and
NDVI, showing that the lagged 3-month SPI had the highest correlation with NDVI
but significant variability existed between months. The highest occurred during the
middle of the growing season, and lowest at the beginning and end of the growing
season (Ji and Peters, 2003). Their regression model revealed significant relationships
between NDVI and SPI in grasslands and croplands, the best occurring in areas with
low soil water-holding capacity (Ji and Peters, 2003). Similarly Liu and Kogan (1996)
showed that NDVI responded quite well to water deficiency with a one-month time lag.
This implies that seasonality and vegetation type must be taken into account when
assessing and monitoring drought with NDVI (Quiring and Ganesh, 2010;
Vicente-Serrano, 2006; Ji and Peters, 2003). In addition, care has to be taken to
account for errors due to data aggregation and/or interpolation across large areas when
using satellite data (Kogan, 1997).
Taking the above into account, in this paper focuses on characterizing
agricultural drought in Karamoja using SPI and assess the suitability of an
78
NDVI-based index to enable spatially explicit monitoring.
4.2.2 Study Area
Figure 4.2: Location of Karamoja Sub-region, showing Meteorological Stations with
fairly complete records. The rainfall gradient decreasing annual average rain-
fall from west to east (Rainfall data source: Hijmans et al. (2005)
The Karamoja sub-region (Figure 4.2) falls within the semi-arid region of East
Africa in North Eastern Uganda between Latitude 1o and 4o North and Longitude 33o
and 35o East. The broader East Africa region is described to have the most impressive
climate anomalies in all of Africa (Trewartha, 1961). According to Baker (1974), the
causes of rainfall variability in East Africa include: diverging and subsiding air streams
that tend to flow parallel with the East African coast, and the failure of cloud
79
formation when dry stable air caps the moist lower air. Rainfall in the region is highly
erratic resulting in recurring droughts that impact livelihoods, resulting in famine,
migration and loss of lives (Stark, 2011; Gartrell, 1985; Baker, 1974).
Most parts of Uganda have bimodal rainfall. Peak rainfall occurs both during
March to May and October to December with annual rainfall close to 1,345-1,500
millimeters (McSweeney et al., 2015) . Karamoja however, has a uni-modal regime
with peak rainfall occurring June to August. The region’s total annual rainfall is
estimated at 500-773 mm with long dry spells and high spatial variability (Egeru et al.,
2014; Government of Uganda, 2007). The rainfall gradient is consistent with the
vegetation transition from the east (driest) to the west (greenest). The wetter and
hence most productive areas include parts of western Kaabong, Moroto, Nakapiripirit,
and the entirety of Abim (Office of the Prime Minister, 2009).
4.3 Data and Methods
4.3.1 Standardized Precipitation Index
The data used in this study including the MODIS NDVI, land use and standard
rainfall are summarized in Table 4.1. Historical daily standard rainfall data obtained
from the Uganda National Meteorological Authority (UNMA) were used to calculate
SPI for the Kotido Station which had at least 30 years of continuous rainfall data,
which is a requirement for calculating SPI following the method outlined by McKee
et al. (1993). Satellite derived rainfall data such as the Climate Hazards Group
80
InfraRed Precipitation with Station data (CHIRPS) (Funk et al., 2015) or WorldClim
(Hijmans et al., 2005) are not used, since these data are calibrated by and interpolated
from meteorological station data. Moreover, satellite measurements are area-averaged
quantities with different spatial distribution and do not capture the variations that
may occur (Hijmans et al., 2005; Flitcroft et al., 1989).
In order to preserve continuity and minimize inaccuracies resulting from missing
data1 in the monthly precipitation time series, linear gap filling was necessary before
aggregating to monthly rainfall time series. To calculate SPI historical data are used to
compute the probability distribution at set timeframes (1month, 2 month, 3 months
etc i.e., monthly and seasonal) observed precipitation totals, and then the probabilities
are normalized using the inverse Gaussian function (McKee et al., 1993; Heim, 2002).
Thus, SPI allows an analyst to measure the rarity of a drought event at a particular
times scale e.g. 1, 2, 3, 6, 9, 12 or 18 months (McKee et al., 1993). Prior to drought
identification and categorization, derived SPI values at 1, 3, 6, 9, and 12 months
(SPI-1, SPI-3, SPI-6, SPI-9 and SPI-12) timescales data were correlated with NDVI
data to determine the most suitable timescale for looking at drought impacts on
agriculture in the region. Generally, short-term SPIs (SPI-1) are used for measuring
meteorological drought while SPI-3 to 6 months are a good measure of agricultural
drought (Zhang and Jia, 2013a). The droughts are categorized based on SPI values as:
mild (0 to -0.99), moderate (-1.00 to -1.49), severe (-1.50 to -1.99), and extreme when
values fall below -2.00 (McKee et al., 1993).
1Data for September to December 1991, May to December 1997 and January to April 2006 were
incomplete.
81
Data Resolution (Period) Source (Reference)
MODIS/NDVI Time Series 250 meters (2000-2011) GLAM (Becker-Reshef et al., 2010a)
Kotido Climate Station Data Daily (1960-2011) Uganda National Meteorological Authority (UNMA)
Croplands Mask 2 meters (2012) (Nakalembe et al., 2017)
Uganda Districts Shapefile Sub-regional (2014) Uganda Bureau of Statistics (UBOS)
Table 4.1: Datasets used in this analysis include rainfall data from Kotido Station and
Karamoja District boundaries subset from the national dataset
Index Equation
Vegetation Condition Index (VCI) (Kogan, 1995) VCI =
NDVI− NDVImin
NDVImax − NDVImin
NDVI Anomaly (Anyamba et al., 2001) ANDVI = NDVI− NDVImean
Standardized Vegetation Condition (SVI) (Ji and Peters, 2003) SVI =
NDVI− NDVImean
σNDVI
Table 4.2: NDVI based indices used in the analysis. Each index was calculated from
NDVI maximum value composite in 14-day time-steps from 12-year (2000 to
2012) time series data.
4.3.2 Normalized Difference Vegetation Index
District level (Abim, Amudat, Kaabong, Kotido, Moroto, Nakapiripirit, and
Napak) 16-day NDVI Maximum Value Composites (MVCs) for Karamoja were
downloaded from the Global Agriculture Monitoring System (GLAM) for East Africa
for 2000 to 2012. GLAM East Africa is an online system for automated processing of
MODIS satellite image time-series and graphs for vegetation condition monitoring and
crop condition monitoring when non-cropland areas are excluded (Becker-Reshef et al.,
82
2010a). These data were used to calculate Vegetation Condition Index (VCI) (Liu and
Kogan, 1996), NDVI-Anomaly and the Standardized Vegetation Index (Table 4.3. The
MODIS NDVI MVCs were retrieved from daily, atmosphere-corrected, bidirectional
surface reflectance using a MODIS-specific compositing method, based on product
quality assurance metrics to remove low quality pixels2 (Vermote et al., 2009). The
MVC pixels contain the best possible observations selected on the basis of
high-observation coverage, low-view angle, the absence of clouds or cloud shadows, and
aerosol loading, retaining the highest-quality NDVI value for each pixel within each
16-day period. The images are spatially continuous and relatively cloud-free
(Becker-Reshef et al., 2010b; Holben, 1986). The data were sub-set using district
polygons from Uganda National Bureau of Statistics (UBOS) 2014 to obtain values
pertaining to each district from the larger MODIS NDVI image. Finally, the 16-day
data were converted to monthly data by averaging two consecutive values pertaining to
each month.
4.3.3 Agricultural land use data
Crop production data were not considered for this study due to data scarcity,
poor quality and inconsistency in reporting for this region. However, in-season and
end-of-season NDVI data combined with cropland spatial extent are used as a proxy
for crop condition and crop failure i.e., agriculture drought. Average NDVI values
within cropland areas identified using a 2012 croplands mask shown in Figure 4.3
(Nakalembe et al., 2017). The croplands mask was developed from very-high resolution
2More information available at modis-land.gsfc.nasa.gov/vi.html
83
WorldView 1 and 2 data acquired between 2010 and 2012 and obtained from the
National Geospatial-Intelligence Agency (NGA) archive through an agreement with
NASA under the NextView License3 (Nakalembe et al., 2017). The crop mask enabled
investigation of drought within agricultural areas by eliminating misleading results
from other land cover types such as pasture and scrubland. Ideally updated land use
data would provide more accurate result these data were not available for this analysis.
To account for seasonality analysis focused at the peak of the growing season
during the months of June, July, August and September (JJAS) when crops and other
vegetation are most sensitive to water availability in Karamoja and the strongest
correlations between NDVI and SPI occur (Figure 4.4). Time-series plots and spatial
maps were derived for the most severe drought years between 2000-2012 at 250m
resolution. Drought years identified through this analysis were cross-referenced with
documentation in official government documents, news and food-aid assistance reports.
3For information about the NextView License, please visit: http://cad4nasa.gsfc.nasa.gov/
84
Figure 4.3: The 2012 croplands map used in this analysis (Nakalembe et al., 2017).
These data show the extent of agricultural land use in Karamoja and the
mean NDVI values within the masked croplands during the month of August
85
Figure 4.4: NDVI temporal profile at Kotido Station (2000 to 2012). The rain season is April to September and NDVI values
are moderate (0.4 to 0.5) during the planting months of April and May. The highest NDVI values (peak greenness)
are during the months of June, July, August and September (0.55 to 0.6)
86
4.3.4 NDVI and SPI Relationships
The relationship between NDVI derived indices (SVI, VCI and ANDVI) and at
different time steps is assessed using simple linear regression. Using a model similar to
Ji and Peters (2003), linear regression with dummy variables (Eq.1) was also done on
ANDVI and SPI-3 to account for the effect of each individual growing season month.
However, in this study the dummy variables are four levels assigned to each of the four
growing season months. Therefore, the regression model containing monthly dummy
variables is expressed as:
ANDV I = β0 + β1(SPI3)
+ β2D1 + β3D2 + β4D3
+ β5D1(SPI3) + β6D2(SPI3)
+ β7D3(SPI3) + 
(4.1)
Where ANDVI is the average NDVI anomaly value within croplands in Kotido
District, SPI-3 is the 3 month SPI at Kotido station ending in June, July, August or
September, D1 - D3 are the dummy variables, β0, β0, .....β11 are regression coefficients,
and  is random error. Dummy variables D1 - D3 are assigned binary values as shown
in Table 4.3;
87
D1 D2 D3
0 0 0 if June NDVI observation
1 0 0 if July NDVI observation
0 1 0 if August NDVI observation
0 0 1 if September NDVI observation
Table 4.3: Dummy Variable Assignment
The regression models corresponding to each of the 4 months are:
JuneANDV I = β0 + β1(SPI3) + 
JulyANDV I = β0 + β1(SPI3) + β2 + β5(SPI3) + 
AugustANDV I = β0 + β1(SPI3) + β3 + β6(SPI3) + 
SeptemberANDV I = β0 + β1(SPI3) + β4 + β7(SPI3) + 
(4.2)
4.3.5 Investigating the relationships between NDVI anomaly and
food insecurity in Karamoja
As a proof-of-concept, the relationship between growing season crop conditions
and food insecurity in Karamoja is investigated. ANDVI (Minimum, Maximum and
Average ANDVI) data are compared with the reported number of people needing food
aid and the population in the different phases of the Integrated Food Security Phase
Classification (IPC) each year. The IPC set of protocols classifies the severity of food
insecurity by drawing from all available information and allows for comparability. Food
88
production is one of the key inputs determining food availability4. The IPC
standardized scale categorizes the severity of acute food insecurity into five levels of
food security (called phases): Minimally Food Secure, Crisis, Stressed, Emergency, and
Famine. Figure 4.5 is an example of the summary assessment for Karamoja while
Figure 4.6) is the summary assessment for Uganda. The underlying assumption is that
in Karamoja crop failure is linearly correlated with food insecurity. Therefore, the
number of people needing food aid will increase with drought severity due to
consequent production losses. Each month’s recorded ANDVI values were used to
estimate the percent of Karamoja’s population needing food aid. The percent of the
population needing food aid each year summarized in Nakalembe et al. (2017), while
the population in each IPC phase were obtained from IPC reports.
4For more information visit http://www.ipcinfo.org/
89
Figure 4.5: IPC assessment summary map for Karamoja (valid November 2015 to May
2016)
90
Figure 4.6: IPC assessment summary map for Uganda (valid September 2009 to January
2010)
91
4.4 Results and discussion
4.4.1 Correlation between rainfall and crop/vegetation conditions
The relationship between NDVI-derived indicators with SPI at 1, 3, 6, 9 and 12
months time steps at the Kotido Station are summarized in Figure 4.4 for cropland
areas during the growing season months of JJAS. Each ending month was correlated
separately for example average June SPI is correlated with June NDVI. Overall SPI-3
had the highest correlations with NDVI data. The ANDVI had the highest and
significant correlation with R2 values between 0.26 and 0.29 for JJAS for all land cover
(results not shown) and improved further to range between 0.31 and 0.32 when
non-cropland areas were excluded.
SPI-1 SPI-3 SPI-6 SPI-9 SPI-12
NDVI Anomaly 0.078 0.324** 0.235** 0.231** 0.187**
VCI 0.103* 0.308** 0.254** 0.241** 0.182**
SVI 0.060 0.316** 0.240** 0.237** 0.196**
N =48
** Correlation is significant at the 0.01 level (2-tailed).
* Correlation is significant at the 0.05 level (2-tailed).
Table 4.4: SPI time scales 1, 3, 6, 9 and 12 months and NDVI based indices at Kotido
Station within cropland areas N=48. The highest R2 values are shown in bold
for each NDVI index. Overall ANDVI data had the highest correlation with
SPI-3.
Figure 4.7 shows that correlations also vary by month that crop conditions are
most influenced during the growing season in Kotido by cumulative rainfall of up to 3
92
months. The highest correlations occurring during the month of August with an R2 of
0.516. Hence, the 3-month timescale is better when using SPI to monitor drought
impacts on crops and vegetation in Kotido. This is because vegetation response is
insufficient at the shorter timescales of 1 or 2 months and the covariation of NDVI
anomaly and SPI reduces after 3 months. August is also the best month for assessing
agricultural drought when using both cumulative rainfall data and ANDVI. It is
interesting to note that Ji and Peters (2003) reached a similar conclusion in a study
assessing vegetation response to drought in the northern Great Plains of the United
States.
93
Figure 4.7: Relationship between SPI (at different time steps) at Kotido Station and NDVI anomaly data for crops only data
2001 to 2012.
94
Variable Coefficient Std. Error t-value Sig.
1. Simple regressiona
Intercept 0.006 0.009 0.689 0.494
SPI3 0.050 0.011 4.692 0.000
2. Regression with dummy variablesb
Intercept 0.006 0.009 0.691 0.493
SPI3 0.034 0.017 2.007 0.051
D1 0.002 0.028 0.054 0.957
D2 0.037 0.028 1.284 0.206
D3 0.046 0.030 1.517 0.137
Table 4.5: Regression analysis on ANDVI and SPI-3 Kotido. aF= 22.012, p-
value< 0.0001, R2=0.324, Adjusted R2=0.309, bF= 6.457, p-value< 0.0001,
R2=0.375, Adjusted R2=0.317
Table 4.6 summarizes the results of the simple regression and regression with
dummy variables tested using ANOVA. Both models were very significant with the
p-value < 0.0001. After eliminating non significant values only predictor variables
SPI3, D1, D2 and D3 are included in the model and the regression models
corresponding to each of the 4 months are:
JuneANDV I = 0.0061 + 0.034(SPI3)
JulyANDV I = 0.0401 + 0.034(SPI3)
AugustANDV I = 0.0426 + 0.034(SPI3)
SeptemberANDV I = 0.0520 + 0.034(SPI3)
(4.3)
The scatter plot in Figure 4.8(a) demonstrates the relationship between observed
95
R2 Linear = 0.324 
• 
0. 1 • • • • ,. • • • 
• • • • 
0.0 • •• • E • • • 0 •• • • • 
> • 
C -
-0.1 • 
• 
• 
• 
-0.2 • 
-0.15 -0.10 -0.05 0.0 0.05 0.10 
Predicted ANDVI using simple regression
(a)
0.1 
0. 0 
0 
<( 
> 
C 
-0.1 
-0.2 
R2 Linear= 0.375 
• 
• 
• 
• 
• 
• 
• 
• 
-0.15 -0.10 -0.05 
• • • • • I:
•• 
• •• • 
• 
0.0 
• • 
• • • 
0.05 0.10 
Predicted ANDVI using regression with dummy variables 
(b)
Figure 4.8: Relationship between SPI-3 at Kotido Station and NDVI anomaly data
for crops only data 2001 to 2012. a) Simple regression, b) regression with
dummy variables
and predicted ANDVI from SPI-3 with simple regression (2001 to 2012). By
accounting for the month effect using regression with dummy variables the relationship
between ANDVI and SPI improved slightly from R2=0.324 (p < 0.0001) to R2=0.375
(Figure 9(b)).
Figure 4.10 shows the temporal relationship of ANDVI and the predicted ANDVI
from SPI with both regressions. There is a clear and significant relationship between
SPI and ANDVI with clear degradation of vegetation conditions or recovery in
response to rainfall. It is important to note that the predicted values for 2010 remain
above zero while the observed values dropped in August. When the two datasets
(ANDVI and SPI) are further analyzed, extreme rainfall at the beginning of 2010
seems to be the cause of this disparity. Observed 3 month cumulative rainfall in 2010
was almost 400% above normal by the end of February and remained above 100%
96
above normal by the end of June. It appears extreme rainfall negatively impacted crop
conditions leading to below average crop conditions by the end of August.
The main conclusion is that NDVI response is most influenced by cumulative
moisture up to 3 months and NDVI anomaly and not de-seasonalized NDVI (SVI or
VCI) having the better relationship at the peak of the growing season during the
Month of August.
4.4.2 Drought identification from rainfall (SPI-3) data
97
 Predicted ANDVI value using regression with dummy variables  •- e-  Predicted ANDVI value using simple regression __•Observed ANOVI
Figure 4.9: June to September (JJAS) crops only observed ANDVI and predicted
ANDVI using simple regression and regression with dummy variables time-
series data for 12 years at the Kotido Station (2000 to 2012).
98
Figure 4.10: August SPI-3 Values 1960 to 2012. The droughts are categorized based on SPI values as: mild (0 to -0.99) in
yellow, moderate (-1.00 to -1.49) in beige, severe (-1.50 to -1.99) in orange, and extreme in red, when values fell
below -2.00 (McKee et al., 1993).
99
Since the highest and significant correlations between SPI-3 and NDVI data
occur during the month of August, in this section droughts in Karamoja are classified
based on the August SPI-3. Between 1960 and 2012 (52 years) extreme agricultural
droughts (shown is red) occurred only in 1965 and 2009, when SPI-3 for the month of
August fell below -2. 1980 was the only severe drought year in the record (orange) and
moderate droughts occurred in 1963, 1979 and 2002 (beige). Mild droughts (yellow)
occurred more frequently (every 2 to 3 years) and occurred in 1960, 1966, 1967, 1968,
1969, 1972, 1982, 1983, 1984, 1985, 1986, 1987, 1988, 1990, 1996, 1997, 1999, 2003,
2005, 2006, and 2007 (Figure 4.10).
By averaging all growing season months SPI-3 values, 1965 was found to be the
worst year on record, with the driest August in the 52-year period followed by 1963
and 2009. August 1979 to August 1988 period marks the longest consecutive period
when drought occurred in the month of August (i.e., when SPI values for the month of
August remained below zero). The period from 1973 to 1978, marks the longest
non-drought period, when SPI-Values remained above average. August 1994 was an
exceptionally good month with the SPI-3 value above 2, which is supported by
exceptional harvests reported during this year. 1994 is named Apamulele by Moroto
district residents, which translates as intact, complete, full or filled up or very good
harvest.
Focusing on the time period 2000 to 2012 (12 years) when SPI and NDVI data
overlap in the study area, 2009 was an extreme drought year followed by 2002 a
moderate drought year. While in 2003, 2005, 2006, 2007 and 2012 mild drought
conditions were experienced during the month of August. Taking a closer look at 2009,
100
the SPI reached a record -3.92 low in April at the start of the season (data not shown),
improving slightly before dropping to -2.16 by August. The next section presents
spatially explicit NDVI data from 2000 to 2012 to explain drought impacts in
agriculture beyond the Kotido Station and to show the differences in spatial severity
across the region.
4.4.3 Spatial and temporal patterns of agricultural drought 2000 to
2012 from NDVI data
Given the relationship between NDVI derived indicators and SPI, ANDVI data
time-series for cropland areas in all other districts in Karamoja are shown in Figure
4.11. It can be observed in these data that there are subtle variations in ANDVI
between districts. However, the major patterns are similar. Overall Abim District
showed the least variation and is the least impacted. While, Moroto District had the
highest inter-annual variability and often the most severely impacted. The differences
between the two districts (indicated with blue arrows) are clear in 2002, 2004, 2007,
2009, 2011 and 2012. 2012 was a year with well distributed and above average rainfall
during all growing season months.
The NDVI Anomaly data also indicate that August 2004 crop conditions were
poorer than August 2006 conditions, which was not captured by the August SPI-3
analysis. We anticipate the extremely low June SPI-3 value (-1.16) had severe impacts
on crops and rainfall in late July was too late to significant ensure recovery of the
crops by the end of August. Therefore, 2004 is re-classified as a moderate agricultural
101
Figure 4.11: The June-September NDVI Anomaly crops only time-series data for Abim,
Amudat, Kaabong, Kotido, Nakapiripirit, Napak and Moroto Districts from
2000 to 2012. The blue arrows highlight the difference in severity between
Moroto and Abim districts.
102
drought.
The derived spatially and temporally explicit NDVI-anomaly maps within
cropped areas at the end of August are shown in Figure 4.12 and further support this
finding. Within cropland areas at least 30% of cropland areas were severely impacted
in 2004, 2005 and 2010, over 75% in 2002 and 100 % in 2009.
103
Figure 4.12: NDVI Anomaly data for the end of August (2001 to 2012)
104
To demonstrate the usefulness of high frequency and spatially explicit data in
monitoring droughts in the region we show in Figure 4.13 the development of drought
conditions in Karamoja from April at the start of season to September for 2009, the
worst drought in 12-year NDVI data record, and 2002. The record lows found in the
SPI-values during April 2009 at the start of season were not evident in ANDVI data
until July 2009 (3 months later) given the slow response of vegetation to water deficits.
During the 2002 growing season SPI-3 data indicated moderate drought conditions
which prevailed to the end of the season. This resulted in crop failure evidenced by
extremely poor crop conditions at the end of the growing season across the region seem
in the ANDVI data.
We also show data for 2008 that we consider a phenomenal year based on the
supplementary data that indicated severe food insecurity. The start of the 2008 season
was poor but SPI-3 values remained above zero. Improvement in vegetation conditions
was seen in the NDVI data but, as indicated earlier this recovery might not have been
sufficient to see crops to maturity since more than 50% of the population required food
aid (Nakalembe et al., 2017). We conclude that the August 2008 ANDVI data are
simply an indicator of general vegetation conditions (grass rather than crops) since
crops are likely not to have survived the poor rainfall during April to June. In
addition, there are no other data to explain the extreme food insecurity in 2008 except
for heightened insecurity in the region during the time. Insecurity could have resulted
in poor production during 2008 and explains the documented food insecurity more
than drought severity. From 2008 to 2011 marks the peak period of the government’s
disarmament campaigns that started in 2004 (Office of the Prime Minister, 2010).
105
Official reports indicate that between 2008 and 2010 more than 50% of the population
required food aid. The assumption is that heightened insecurity disrupted day-to-day
activities; crop cultivation was perilous in vast areas of the region and irrespective of
good rainfall, harvests were very poor.
The observations above demonstrate part of the usefulness of combining
meteorological and satellite-based indices when understanding drought severity for
better drought monitoring (Zhang and Jia, 2013a). Without the spatially-explicit
NDVI information classification of droughts, simply analyzing rainfall anomaly would
not provide the extent and differences in severity of drought across the region. For
example, without the NDVI information, extreme rainfall would indicate a good year,
however this is not necessarily the case since heavy rainfall often leads to flooding that
physically damages crops (e.g. in 2010). This information; where, when and how
severe drought impacts are, is important when planning response e.g., when
determining if, where and when to begin providing humanitarian assistance.
106
April 2002 May 2002 June 2002 July 2002 August 2002 September 2002
April 2008 May 2008 June 2008 July 2008 August 2008 September 2008
April 2009 May 2009 June 2009 July 2009 August 2009 September 2009
-1 -0.5 0 0.5 1
Mild Moderate Severe Extretne 
Figure 4.13: End of month 16-day NDVI Anomaly maps from May to September for 2002, 2008 and 2009
107
4.4.4 Relationship between NDVI and number of people requiring
food aid
The main goal of this section was in a proof-of-concept frame use ANDVI data to
estimate the impact of agricultural droughts on food security in Karamoja. It is
important to note that the relationships presented below are derived from a very small
sample, and it assumes there is a relationship between vegetation condition, crop
production and domestic food supply. Nevertheless this would be a quantitative, easily
reproducible and quick method to estimate food aid need in this region. This simple
and straight forward model used area averaged seasonal ANDVImin, ANDVImean, and
ANDVImax to estimate the percent of the population likely to need food aid each year
(equation 3) and the percent of the population likely to fall in the IPC Phase 3 (Crisis)
by the end of the year (equation 4) .
The regression model equations are shown below. Future studies will continue to
evaluate this model as more data become available for more years.
IPC Phase 3 (percent of population) = 87.47− 358 ∗ ANDVImax (4.4)
Food aid need (percent of population) = 55.17− 476 ∗ ANDVImax (4.5)
In spite of the current limited sample size (n=5) correlations between ANDVI
and food aid need were high. ANDVImax had the highest correlation with the R
2=
0.907 (Table 4.6). This implies that 90.7% of the variation in the number of food aid
needed can be explained by seasonal ANDVImax. Similarly, correlations with ANDVI
and population in IPC Phase 3 (n=8) were high. Irrespective of seasonal ANDVI value
108
ANDVIvalue Food aid need IPC Phase 3
r-value 0.724 0.724*
ANDVImin R
2 0.524 0.524
RMSE (%) 9.9 19.3
r-value 0.881* 0.730*
ANDVImean R
2 0.777 0.533
RMSE (%) 6.8 19.1
r-value 0.952* 0.708*
ANDVImax R
2 0.907 0.501
RMSE (%) 4.4 19.7
Table 4.6: Pearson Correlation Coefficients (r) between ANDVI (min, mean and max)
values and food aid need (n=5), and IPC Phase 3 (n=8). Bold r-values
indicate cases with highest correlation coefficient.*Indicate values significant
at the 0.05
use (min, mean or max) at least 50% of the variation in the population in IPC Phase 3
can be explained by ANDVI.
From this proof-of-concept estimate, at least 72% of the population in Karamoja
(about 759,240 people5 will require food aid to make through to the 2017 harvest
Figure 4.17(a). The model also estimated that at least 30% (316,350 people) of the
population would fall in the IPC 3 (Crisis phase) by December 2016 (Figure 4.14(b)).
Similarly, the model demonstrated that it possible within a reasonable margin of
error (less than 10%) to estimate the population that will likely need support as early
5Estimated from UBOS 2016 population projection
109
as August (at the peak of the growing season). We hope that as more monitoring data
on food security for Karamoja become available, better early warning, leading to
informed decision using remote sensing can be realized.
110
(a)
(b)
Figure 4.14: (a) Observed and projected (from ANDVI) number of people needing food aid and (b) reported and projected
(from ANDVI) number of people in IPC Phase 3.
111
4.5 Concluding remarks
The lack of clarity concerning occurrence, boundaries and possible impacts of
drought on agriculture remains an obstacle to agricultural development in the region.
This information is now more critical with the current and projected rates of extreme
events in this region (Cervigni and Morris, 2016).
Agriculture in Karamoja is nearly 100% rain-fed and the soils have poor
water-holding capacity resulting in high sensitivity of crops to drought conditions
(Nakalembe et al., 2017). As a result, and with no irrigation, consistent rainfall during
the growing season is necessary for above average production in the region. Rainfall is
often poorly distributed, which explains the region’s wide growth of sorghum, a
significantly drought-tolerant crop. However, its performance has not been sufficient to
ensure food security as unpredictable shocks often decimate the crop sometimes
causing complete crop loss.
The majority of farmers in Karamoja maintain their cropping calendar and once
seeds are sowed, many are not in a position to replant if rains fail at the start of the
season. Extended dry periods often result in extreme crop failures irrespective of a
good start of the rainy season.
In this study demonstrated the suitability of MODIS NDVI data for monitoring
and characterizing agricultural drought in Karamoja and shows that the
satellite-derived data provide sufficient details for spatially-explicit drought assessment.
The SPI provides good point measurements but in this region it is of limited use due
to the scarcity of climate station data, leading to the lack of spatial information.
112
Though without extensive testing, we have shown that it is possible to predict
the number of people who require food aid during the lean season in Karamaoja
(December to March) within a reasonable margin of error at the peak of the growing
season. With the sparsity of information, this method could help in planning for timely
and appropriate response.
This study also highlights the need for continued detailed assessment of drought
to minimize over generalization and misleading information. Assessments that lack
empirical backing have to be used with extreme caution in decision-making. Apart
from detailed spatial and temporal drought information, more research is needed
regarding drought impacts in Karamoja. Human perceptions of drought in the region
are necessary to help understand additional underlying causal factors of severe crop
failure (e.g. agricultural practices). In addition, other factors including pests, flash
floods and wind erosion can cause substantial crop damage from year to year; these
need to be considered when examining the suitability and sustainabilityagricultural
production.
There is a general lack/poor distribution of meteorological data that would
otherwise form the basis for the design of crop insurance programs to enable a timely
and appropriate response to help farmers (Turvey and Mclaurin, 2012; Rowley et al.,
2007). There is some indication that NDVI data can be fill this gap. This has been
demonstrated in Kenya using VCI Klisch and Atzberger (2016) and similarly Uganda
is in the process of designing its Disaster Risk Financing Program using MODIS NDVI
data6.
6As per NUSAF 3 Disaster Risk Financing Handbook
113
There is also a strong need to educate farmers, and to provide early
warning/forecast information to support decision-making. Most importantly there is
need for behavioral change. If farmers are provided information early, real positive and
informed change can be realized including, changing planting dates, adopting drought
adapted crop varieties and irrigation to realize improvement in crop production.
The Intergovernmental Panel on Climate Change (IPCC) projects an increase in
extreme events such as drought in the sub-region. Observations of rainfall for Uganda
(1960 to 2010) show significant decreasing rainfall trends for March, April, May
(MAM) and June, July, August (JJA) a period that covers the entire growing season in
Karamoja (McSweeney et al., 2015). This is coupled with an increase in temperature
at an average rate of 0.28oC per decade (McSweeney et al., 2015). Uganda is already
experiencing the negative impacts of climate change witnessed in the reduced in rainfall
amounts, and according to (Funk et al., 2012) Karamoja is projected to experience
further decrease in June to September rainfall of 20 millimeters and temperature
increase of greater than 1.5oC from 2010 to 2039. This implies that drought frequency;
severity and impacts are likely to increase making it paramount to have a functioning
drought monitoring system to minimize negative impacts on the people of Karamoja.
This study shows that satellite data can provide the much-needed data to fill the
data gap that inhibits long-term drought monitoring at a significantly lower cost than
traditional climate station-based monitoring in data scarce regions like Karamoja.
With more spatial details, satellite data provides a baseline for understanding
long-term food insecurity in the region. Satellite-derived data from decision support
systems such as the GLAM system used in this study can also be used near real-time
114
drought monitoring. This would enable proactive assessment, planning and response,
and would limit incidences where information gets to the decision-maker too late. The
country would be able to avoid incidences such as that of the 2015 drought, when the
Karamoja region experienced famine and no action was taken until the food security
situation had deteriorated in September NewVision (2015), yet extreme crop failure
was eminent earlier and was evident in MODIS NDVI by early July 2015 (Office of the
Prime Minister, 2015).
115

5 Human Perceptions of Drought in
Karamoja Uganda: a Moroto
District Case Study
5.1 Summary
Understanding human perceptions of drought can facilitate better interventions
that are evidence-based. This chapter combines the results from interviews and
observations from remotely sensed information to document and assess human
perceptions of drought in Moroto District in the Karamoja sub-region of northeastern
Uganda. Although drought exposure, sensitivity, and impacts are evident, local human
perceptions of drought are not well understood in Karamoja. Results from this study
indicate that past experiences of drought do not necessarily influence expectations of
future drought. From the interviews conducted in this study, it appears that majority
of farmers are entrenched in poverty, extremely vulnerable to drought events and
depend on emergency food assistance for survival. Furthermore, farmers do not feel
117
empowered to plan for and manage drought impacts through agricultural practices.
Traditionally, the Karamojong were pastoralists that were settled by the Government,
with farming as an intended livelihood. This study shows that factors such as armed
conflict and external interventions intermingle with traditional culture to have
profound direct influences on farmers’ perception of drought amongst communities in
Moroto District. External to farmers’ practices, these factors continue to disrupt
communities, limiting development and adoption of better agricultural and drought
management practices. This work also illustrates that the perception of drought by
un-empowered farmers experiencing one ’extreme’ year after another, are distinct and
pose greater challenges for policy-making. A situation where farmers are unaware of
the risks they face and/or assume drought management is entirely the government’s
responsibility is particularly challenging, leading to disconnect between perceptions,
decision and actions taken. Because the majority of the population (over 75%) is
extremely poor, simply limiting or reducing access to food aid will not likely yield the
desired outcomes. Similarly, new technologies alone (such as irrigation) will not
provide the solution. A unique set of programs has to be developed to encourage and
help farmers to develop their adaptation strategies that can then be supported with
appropriate technologies. To be effective, development agendas for the region need to
be informed by local farmers’ perceptions and encourage farmer-led risk assessments,
and adaptation and mitigation strategies, based on practical experience working with
the community.
118
5.2 Introduction
Rainfall is the dominant driver of ecosystem dynamics and the major constraint
on human land use in pastoral areas in East Africa (Ellis and Galvin, 1994). The
region experiences frequent severe droughts, chronic food insecurity, and continuous
internal and external conflict. Furthermore, it is largely entrenched in poverty, faces
immense external pressure, and is often the center of humanitarian crises. Drought
impacts in semi-arid East Africa today are in part due to increased population and
ill-suited farming practices on marginal lands (Nakalembe et al., 2017). In recent years,
harsh climate conditions have led to greater competition for food, water and pasture
(Becker-Reshef et al., 2010a; McCarthy et al., 2001). Moreover, East Africa is projected
to face more adverse impacts with projected climate change. For example, a projected
increase in daytime temperatures above 30oC will lead to a 10 to 20% decrease in
yields of rain-fed crops, including maize (Kassie et al., 2013; Wolfram and Lobell,
2010). In general, farmers in semi-arid East Africa perceive drought as the primary
factor in reducing production and understand that increasing climate variability is
likely to exacerbate the damaging effects of recurrent droughts (Fisher and Snapp,
2014). However, the social and economic impacts of future droughts on agriculture in
the region will depend on the local farmers’ adaptive capacity and how they perceive,
prepare for and respond when droughts occur (van Duinen et al., 2015; Slegers, 2008).
Understanding human perceptions of drought can contribute to our
understanding of the behavioral responses of the communities at risk (Kassie et al.,
2013; Dagel, 2010; Taylor et al., 1988). Human perception is the range of judgments,
119
beliefs, and attitudes that underpin human decisions and choices (Taylor et al., 1988).
Decisions and responses are ideally manifested in adaptation and mitigation measures
that have a direct impact on the environment and influence changes in the landscape.
For example, one way of managing drought impacts would be changing the cropping
calendar and plant according to the new calendar’s timeline. A non-responsive action
would be to continue to plant traditional seed varieties not suited for shorter rain
seasons. Studies of environmental perception show that behavior is influenced by
subjective images of the environment, attitudes, goals, feelings and beliefs and these, in
turn, affect farmers’ actions in preparation for future events (Howe, 2009; Taylor et al.,
1988; Bunting and Guelke, 1979). In other words, personal experience of disaster
should lead to higher risk perception (Diggs, 1991; Slegers, 2008). Thus, for people to
adapt successfully, they must accurately perceive their relative risks to develop a better
heuristic for anticipating hazards and taking precautionary measures (Howe, 2009). An
example is changing planting dates in response to consecutive seasons of delayed onset
of rainfall.
However, the relationship between perception and behavior is not simple or
direct, and actions depend upon the perceived range of alternative measures and the
information available to the person (Taylor et al., 1988). In the long run, perceptions
impact on a community’s sensitivity to drought and their capacity to adapt and should
be acknowledged and taken into account by policy-makers. Understanding the
formation of drought risk perception should be a prerequisite to designing productive
and more resource-efficient drought risk management strategies (van Duinen et al.,
2015; Howe, 2009; Adger et al., 2007).
120
Greater perception of risks can lead to effective precautionary actions (Burton,
1963). However, studies of farmer households in Sub-Saharan Africa have shown that
farmers often over estimate risks associated with rainfall variability, for example Kassie
et al. (2013), Gandure et al. (2013) and Mubaya et al. (2012). While in a study of
farmers in Goima, Tanzania, Slegers (2008) documented an underestimation of drought
risk because farmers in the region regularly face partial crop failure. In both cases,
poor estimation of risks can lead to the adoption of sub-optimal adaptation and
mitigation measures. However, these studies also document quite extensive positive
associations between farmers’ perceptions of drought and climate risk and locally
developed adaptation strategies to minimize production losses (Gandure et al., 2013;
Kassie et al., 2013; Mubaya et al., 2012; Slegers, 2008). For example, in southern
Malawi, farmers that value crops that mature early and are drought tolerant were
found more likely to adopt and continue to use drought resistant maize (Fisher and
Snapp, 2014). While in Central Rift and Kobo Valley in Ethiopia, Kassie et al. (2013)
document changes in farming practices including adjustment of the cropping calendar
and in-situ water conservation that have been beneficial to farmers in avoiding
complete crop failure.
Wealth plays a significant role in determining which households are permanently
hurt by drought and climatic variability. Predominantly, it is the poorest that lack
access to off-farm income that are highly exposed to agro-climatic risks (Gandure
et al., 2013; Lybbert et al., 2009; Reardon and Taylor, 1996). Wealthier households
tend to smooth consumption by reducing income spent during drought periods while
poorer households tend to reduce consumption and dispose of buffer assets when
121
lower-cost options have been exhausted (Lybbert et al., 2009). Hence, the actions of
wealthier farmers may reflect better perceptions than those of the poor who lack the
means to take mitigation or management actions. A fundamental question that arises,
therefore, is how do farmers with no buffer assets manage drought’
Precautionary action should increase with the severity of damage experienced.
Furthermore, a better understanding of farmers’ perceptions of current and future
droughts can provide policy makers and development planners critical information
needed for better interventions that are more suited to the context of existing practices
(Kassie et al., 2013; Howe, 2009). However, despite the obvious central role of local
communities (farmers and pastoralist) in land-use, production and development in
Karamoja, their perceptions are rarely considered in development programming for the
region. It is this study’s underlying assumption that the lack of basic understanding of
the decision making processes of local communities is detrimental and limiting to
development programming in the region. The primary objective of this study is
therefore to assess Karamoja farmer’s perceptions of agricultural drought to uncover
some of the potential limitations and misapplications of current agricultural
development programs. Distinctive in this study, is the documenting of perceptions of
a society facing complex development challenges that are rooted in extreme poverty1 ;
a lack of a cultural farming tradition; high frequency of drought events; systemic
dependence on food aid; a long history of violent conflict in the form of cattle raiding
that aggravates the other factors; and the external government forces that have placed
1Poverty Headcount ratio was 75% in 2012/13 with mean monthly income per adult approximately
$10 per month
122
them in this situation.
5.3 Materials and Methods
5.3.1 Study Area
As shown in Figure 5.1, Moroto District is located in Karamoja sub-region of
northeastern Uganda between 1o53’N and 3o05’ North latitude and 33o38’ and 34o56’
East longitude. Moroto has a total land area of 3,532 km2 and is bordered by Kaabong
in the north, Kotido in the northwest, Napak in the West, Nakapiripirit in the
Southwest, Amudat District in the South and the Republic of Kenya to the East.
According to the 2014 National Census, Moroto District had a population of 103,432,
with a 2.4% growth rate. Population density is very low at 29 person/km2 when
compared with other districts in the region, for example Kotido is at 50 person/km2
(Uganda Bureau of Statistics, 2016). Moroto district is semiarid with the dry season
from November to March. The district has a mean maximum temperature of 28oC.
The hottest months are January and February with an average maximum temperature
of 33.5oC. The rainy season begins early March and lasts until August, and annual
rainfall ranges between 300mm and 1200mm with mean annual rainfall of 800mm
(Figure 5.2). During the rainy season Moroto district experiences flash flooding, that
mainly affect areas surrounding streams and rivers. Flash floods often last a few hours
after heavy down pour often causing severe erosion in low-lying areas, which damages
roads making neighboring areas inaccessible.
123
Figure 5.1: Location of study area in Uganda
Savannah grasslands with scattered trees dominate the area. The district’s forest
cover is estimated at 100 km2. Forests are mostly found on hills and on Mt. Moroto.
Mountain Moroto located within the district boundaries is an extinct volcano covering
483 km2 from 960 to 3084m altitude. The mountain has a wide range of vegetation
ranging from dry savannah woodland to closed canopy montane forest above 2000m,
and an estimated 203 species of trees and shrubs (Howard et al., 1997). The people of
Moroto are predominantly agro-pastoralist practicing both subsistence agriculture and
semi-nomadic livestock rearing. Moroto District is therefore, dominated by two
livelihood zones: the Karamoja Mountain and Foot Hills Maize and Cattle Zone
(UG04), mostly in Tapac sub-county; and the Central Karamoja Sorghum and
Livestock Zone (UG05) (See Figure 2.4 in Chapter 2). The main crops grown include
124
Figure 5.2: Average rainfall for Moroto District. Rainfall data from CHIRPS Data and
NDVI Data from GLAM East Africa
sorghum, maize, beans, cow peas, sunflower, groundnuts, and vegetables. In addition
to severe crop losses due to sporadic rainfall, crop diseases and pests often cause
extensive damage in Moroto.
People mainly live in clusters locally known as Manyattas that are fenced off
with thorny shrubs. Fifty to four hundred people usually occupy a Manyatta. Some
villages are gradually depopulated during the dry season when pastoralists migrate to
refuge grazing areas. Over-dependence on natural resources has led to extensive
degradation of the natural environment in the District. Due to erratic and unreliable
rainfall, the leading cause of crop failure, many communities engage in alternative
economic activities, most of which are detrimental to the environment. Increasingly,
entire villages are dependent on tree cutting for building material, fuel and charcoal
burning. Sub-counties such as Nadunget and Rupa have been profoundly affected by
such practices. Currently, these Sub-counties have large tracts of land that are likely to
125
remain bare, with deeply eroded gullies.
Like most of the subregions of semi-arid East Africa, the frequency of
drought-induced famine has increased over the past 30 years in Karamoja. Estimates
indicate that severe and even climatically mild droughts lead to severe crop failure in
the region leaving over 50% of the population in need of emergency assistance
(Nakalembe, 2017). Drought induced famines often result in loss of human life and a
complete breakdown of communities in Karamoja. Often, the elderly and children die
from starvation and severe malnutrition despite continued flow of aid and development
projects in the region.
5.3.2 Data Collection and Analysis
This study was based on a baseline questionnaire, in-depth interviews with
farmers and key informants, and satellite remote sensing information. Preliminary field
research began with a baseline survey in August 2013 after approval from the
Institutional Review Board (IRB) at the University of Maryland2 and from the
Uganda National Council for Science and Technology (UNCST)3. Interviewee selection
followed a purposeful sampling approach following referrals of Community
Development Officers (CDOs), and District Agricultural Officers (DAOs).
The interviews lasting 45 to 60 minutes were intended to obtain information on
general farming practices in the area and to validate and identify potential survey
topics on drought, agriculture and land use in the district. After visiting the district
2IRB Project Ref: 363568-2 (25 June 2013) Appendix 1 and 363568-3 (13 March 2014) Appendix 2
3Research Approval Ref: A 473 (16 May 2013) Appendix 3
126
and sub-county offices and several households in Moroto, semi-structured interviews
were conducted with farmers at their homes with the support of county community
workers. Only household heads were interviewed. Conducting questionnaire-based
research in Moroto can be an arduous task. On occasion, it was impossible to complete
the questionnaire as respondents abandoned the interviews and walked away.
Sometimes we faced hostility when community members thought we were
land-grabbers. Following this experience, arrangements were made with district officers
to identify farmers that would be interviewed in a more conducive environment with a
smaller, carefully selected sample. The sample size was not meant to be representative,
but rather to maximize insight and understanding. With careful identification, a small
sample could be used to generate significant and reoccurring themes (Marshall, 1996).
In 2014, semi-structured interviews were conducted at three sub-county offices
with a sample of fourteen men and seven women farmers. The goal was to maximize
an in-depth understanding of challenges faced by a small group rather than a
superficial knowledge of a larger group. The sample was also limited due to time
constraints and accessibility to farmers and resources. The in-depth interview sample
comprised farmers from Nadunget, Rupa and Katikekile sub-counties and fourteen
villages: Namorotomei, Singila, Nakiloro, Nadiket, Acherer, Lotirir, Nakiloro,
Natumkaskou, Lorukumo, Kadilakeny, Namogorat, Lokutimo, Lonyasan, and
Kalukalet. No farmers were identified in Tapac sub-county due to security concerns.
Most villages in Tapac located in Mount Moroto, are inaccessible and at the time
locals were considered armed and dangerous.
The surveyed farmers ranged in age and experience. Questions focused on the
127
farmers’ memory of past droughts, or years that were particularly severe. The farmers
were also asked both broad and general questions about their experiences of past
droughts, followed by a discussion of their expectations of future occurrences. The
interviews were recorded in the local dialect and conveyed in English by the sub-county
CDOs and DAOs .
In this study, farmers’ perceptions of drought in Moroto District are presented
following the critical elements of experience, memory, definition, and expectation
(Taylor et al., 1988; Diggs, 1991; Slegers, 2008). Similar to Taylor et al. (1988) (Figure
5.3) the results are summarized and describe as derived meanings from individuals
experiences of drought in Moroto. For this study, identification of drought years
(experience) was achieved by analyzing the joint variability of the Standardized
Precipitation Index (SPI) and satellite-based Normalized Difference Vegetation Index
(NDVI) anomaly data in Karamoja region (Nakalembe, 2017). Memory, definition, and
expectations are written as composite descriptions focusing on common experiences
from transcripts of in-depth interviews conducted in 2014. Table 5.1 summarizes the
characteristics of the surveyed farmers in Moroto District and the drought years they
recalled.
As part of the process of analyzing the interview material, emerging themes were
categorized and coded from each of the responses to the guiding interview questions.
Each grouping was repeatedly reviewed within the context of the participant’s
complete answers to overall questions to ensure no response was taken out of context.
The groupings were organized under each central guiding question, taking note of
newer themes that developed from the data. Table 5.2 summarizes essential thematic
128
Figure 5.3: Elements of drought perceptions (Taylor et al., 1988) extracted from
(Slegers, 2008)
statements and their frequencies in response to the guiding question one. The
frequency refers to the number of times participants made the same or similar
statement.
Data from the interviews were transcribed, organized, and read multiple times.
Quotes that provided a clear understanding of participants’ perceptions were
highlighted and selected as illustrative. The essential thematic statements were based
on the participants’ descriptive expressions, and illustrative quotes were chosen based
on how well they captured the experience or memory in question. Farmers’ memories
were also compared to an independent assessment of drought derived from time-series
satellite-based NDVI data.
5.4 Results
5.4.1 Drought experience and memory
Drought occurrence and degree of severity (experience) were measured by the
3-months Standardized Precipitation Index (SPI-3) and NDVI data analysis from
129
Form ID Sub-county Age Sex Farming experience Drought Years Recalled Worst Drought
6 Rupa <30 Female <5 2014, 2011 2014
1 Rupa <30 Male <5 2014, 2002, 2004, 2009 2002
2 Rupa <30 Male <5 2014, 2009, 2007, 1992 2007
1 Nadugnet <30 Female 5 to 15 years 2013, 2011,2007
7 Katikekile <30 Female 5 to 15 years 2014,2004, 2005 2004
1 Katikekile <30 Male 5 to 15 years 2009, 2014 2009
3 Katikekile <30 Male More than 15 years 2014, 2011, 2012, 2001 2011
5 Katikekile <30 Male More than 15 years 2014,2009,2011, 2005 2009
7 Rupa 31 to 60 Female More than 15 years 2014, 2012, 1985 1985
8 Rupa 31 to 60 Female More than 15 years 2014, 2005, 1980, 2009 1980
4 Rupa 31 to 60 Male More than 15 years 2008, 2004 2008
2 Nadugnet 31 to 60 Male More than 15 years no sure, 1984 1984
6 Katikekile 31 to 60 Female More than 15 years 2013, 2014 2013
2 Katikekile 31 to 60 Male More than 15 years 2014, 2013,2013, 1992, 1993,2012 2012
4 Katikekile 31 to 60 Male More than 15 years 1991, 1994, 2008, 2001 1992
3 Nadugnet 60+ Female 5 to 15 years 1980, 2001, 1984 2002
4 Nadugnet 60+ Male 5 to 15 years 2012, 2014, 1984 1984
3 Rupa 60+ Male More than 15 years 2004, 2013, 2009, 1982 2008
5 Rupa 60+ Male More than 15 years 1980, 2009, 2014 1980
5 Nadugnet 60+ Female More than 15 years 1984, 2014, 2011 , also every year since 1994 1984
6 Nadugnet 60+ Male More than 15 years 2007, 2000, 2002, 2014, 1984 1984
Table 5.1: Characteristics of the surveyed farmers in Moroto District
130
Guiding Question Thematic Statement Frequency
What have you experienced in terms of drought? Lack of water 75
Disease outbreaks 71
Too much sunshine 50
Lack of food (Famine) 48
Lack of pasture 45
Deaths 41
Migration 33
Conflicts 13
Poverty 13
Malnutrition 11
Increased prices for food 11
Theft 10
Increased suicide cases 5
Table 5.2: Main thematic statements from interviews under guiding question 1
131
Chapter 2. Drought memory in this study consists of those events that were part of
the farmers’ direct experience and recollection (Nakalembe, 2017; Taylor et al., 1988).
Figure 4 illustrates recorded meteorological droughts (bars), normalized NDVI data
(black line), and drought years recalled by farmers since 1980 (denoted by black dots).
From the climate record, 2009 was the only exceptional drought year (shown in red),
when the SPI-3 fell below -2 (Nakalembe, 2017). 1980 was the only ’extreme’ drought
year on record (brown), and 2002 was a ’severe’ drought year (pink). Moderate
droughts were experienced during 1984, 1987 and 1990 while mild droughts (yellow)
occurred more frequently footnoteMild droughts occurred in 1982, 1983, 1985,
1986,1996, 1997, 1999, 2003, 2005, 2006, 2007 and 2012 (Nakalembe, 2017).
The cumulative analysis of livelihood risks (Table 2) shows that water scarcity
poses the highest risk to farmers in Moroto. Disease outbreaks are the second most
important factor followed by extreme temperatures (too much sunshine). Lack of food
and pasture rank third and fourth respectively, followed by deaths, migration, and
conflict.
As expected, older farmers have experienced more droughts. All farmers older
than 60 years recalled droughts from the 1980’s, while only 43% of farmers aged 31 to
60 mentioned droughts from the 1980s (Table 5.3) It is also interesting to note that all
farmers recalling droughts from the 1980s indicated that these were the worst droughts
to their recollection. The year 1984 was referred to as the worst drought period. Most
farmers recalled that many people died of hunger that year. From 1980 to 1990 marks
the longest drought period except 1981 and 1989, which were non-drought years
132
F
ar
m
er
s 
M
en
ti
on
in
g 
Y
ea
r
Figure 5.4: Severity of droughts in Karamoja 1960 to 2012. August SPI-3 Values 1980
to 2012. The droughts are categorized based on SPI values as: abnormally
dry less than 0 to -0.7, moderate (-0.80 to -1.2) in pink, severe (-1.3 to -1.5)
in orange, extreme (-1.6 to -1.9) in brown, and exceptional in red, when
values fell below -2.00 (McKee et al., 1993).
133
“The drought that I still remember and cannot forget occurred in 1984, It
started during change of government... food items disappeared from the
market, there was an outbreak of diseases, which claimed both animals and
people, too much sunshine, strong winds, and displacement of people. Most
people from Matheniko went to Namalu and later came back in 1990. Also
in the in 1980s many people died and many migrated to Naoi parish and
Regina mundi, Goli Army barracks. We received relief including cooking oil,
powder milk, biscuits and other food items from the community center.”
It was rather surprising that only 33% of farmers recalled the 2009 drought,
which from both meteorological and satellite data records was the worst on record,
with the SPI values falling below -2 (exceptional drought). When specifically asked
about 2009, 100% of the farmers indicated that 2009 was indeed a drought year, but
the impacts seem to have been varied in the District. For example, two farmers (one
from Rupa and another from Nadunget) indicated that they were able to harvest their
crops in 2009 and losses were not so severe. Further analysis of NDVI data shows that
vegetation conditions were extremely poor (the worst on record since 2000) in all three
sub-counties, which contradicts the farmers’ recollections. Figure 5.5 shows NDVI time
series data for the three sub-counties indicating severe drought conditions by 19 July
to 7 September and the season remained the worst on record since 2000.
A total of 18 (86%) of the farmers indicated that they were currently (August
2014) in drought, although SPI-3 data indicated normal conditions. The NDVI data
for Moroto District showed slightly below average conditions at the time of the
134
Figure 5.5: Average vegetation conditions from NDVI data in Nadunget, Katikekile and
Rupa sub-counties in 2009 compared to long-term average, maximum and
minimum NDVI values between 2000 and 2016. Average conditions shown
similar trends in all three sub-counties with clear deterioration to the worst
on record by July 19
Farmer Identifying drought during Raining less
Median age No. 1980s 1990s 2000s Yes (%)
Under 30 8 0% (0) 12% (1) 38% (8) 100
31 to 60 7 43% (3) 28% (2) 33% (7) 100
60 and over 6 100% (6) 16%(1) 28% (6) 100
21 42.8% (9) 19% (4) 100% (21) 100
Table 5.3: Age and experience of drought by farmers in Moroto District
135
interviews. In addition, photos taken during that period show poor conditions that
were attributed to the delay of the rainy season. Between 2000 and 2014, at least one
farmer indicated each year as a drought year. Except for the years 2009 and 1984, no
meaningful relationship was found between the actual occurrence of a severe drought
year and the memory of it, since many non-drought years were also recalled as drought
years. The excerpt below from an interview with a key informant at the district level
offers insight into farmers’ experiences during drought years in Moroto.
“The prolonged dry spells tend to be the end of life in Moroto since dry
spells affect farmlands and animals that are essential components of
survival for the Karamojong. Due to unreliable rainfall and prolonged dry
spells, crops fail to germinate forcing farmers to replant over and over
again. After investing so much in their gardens, such as buying seeds, hiring
ox-ploughs and casual laborers, some farmers get frustrated with the
thought of losing all that money. It is wasted energy and resources when
you plant crops, and they do not germinate. Then you have to buy more
seeds and sow again. After several failed attempts farmers eventually give
up on cultivating their gardens. During a famine, people sell whatever is
available to them in the effort of making money for survival. Normally,
items like building materials and grass are freely available, but during
famine periods, these are sold at high prices. Those who own property such
as land and animals are forced to sell them to be able to afford the
expensive food available at local markets. Selling of animals is not
136
profitable due to low prices that are offered by buyers if the animals are not
healthy. Some animal owners try to keep their animals as they wait for the
weather to change so that they can feed them and make some profit. People
are forced to eat their sick animals instead of letting them die.”
5.4.2 Farmer definitions of drought
Many drought definitions exist, and there is no universal or precise definition of
drought today (Ji and Peters, 2003). Droughts mean different things to different
people. To understand the potential impacts of drought on livelihoods, definitions of
drought must identify regional differences in terms of specific climatic conditions,
physiological characteristics, economic development, and traditions (Taylor et al., 1988;
Kogan, 1995). From this study, 66% of the farmers defined drought in terms of
unreliable rainfall and a lack of water, and 40% mentioned too much sunshine (high
temperatures). In other words, droughts are identified as hot and dry periods in
Moroto. Most interviewees (85%), associated drought with social and economic
impacts such as disease and famine; (71%) mentioned human and livestock deaths; and
(71%) highlighted the increase in migration. Other consequences included increases in
crime, which mostly involved thefts, spikes in domestic violence and divorces. Before
the completion of the disarmament campaigns, violent raids would spike during
drought periods, often leaving entire villages destroyed (Nakalembe et al., 2017).
Rainfall unreliability was a recurring theme, and many farmers pointed out that, dry
spells lasting from a few days to several weeks often result in complete crop failure.
137
According to one farmer from Rupa Sub-county, Natopojo village (Farmer ID
No2),
“When there is a drought, there are strong dry winds that continuously
blow from east to the west. In 2009 in Nakiloro, three-quarters of the
animal population died as a result of famine. Also in 2007 and 2008 most
animals fell sick and died due to the outbreak of diseases. In 2006, members
of 5 families died in Lope Sub-County in Napak District due to famine. In
1999 two people (Kite and Lower) were killed at Monariwon, when they
participated in a raid as a means of sustaining their families.”
Surprisingly, many mild and non-drought years4 from the SPI assessment were
also recalled by farmers as drought years, an indication of a low severity definition. A
low severity definition is determined when mild droughts and climatologically
insignificant dry spells are defined as drought years by farmers (Taylor et al. (1988).
However, farmers’ recollections can be supported by the very high drought sensitivity
of the region (see. Nakalembe, 2017b). Karamoja is highly sensitive to drought, and it
is not unusual for short dry spells and mild droughts to severely affect crops without
the possibility of recovery. The above was observed first hand in August 2014 when a
short-lived but severe dry spell destroyed crops. The dry spell was followed by heavy
rainfall that resulted in pasture recovery during September 2014 (see Figure 5.6)but
was too late for crops to recover. Which explains why 85% of the farmers that were
interviewed indicated that they were in a drought at the time.
4Mild and non-drought years included; 1991, 1993, 2000, 2001, 2004, 2008, 2011, 2013 and 2014
138
Figure 5.6: Rainfall anomaly data for Moroto District shows extreme dry conditions
until the 3rd dekad of August.
5.4.3 Expectations of drought
According to one District Officer, chronic droughts and unpredictable weather
over the years have tremendously changed the traditional agricultural calendar that
farmers in Karamoja used to follow. All interviewed farmers indicated that not only is
it raining less, but that, the little rainfall they get has become very unreliable. When
asked if droughts were getting worse, 66% said that droughts were not getting worse as
opposed to 24% who felt droughts were getting worse; 2 farmers (9%) were not sure.
All farmers from Nadunget indicated that droughts were not getting worse but more
frequent, and some did identify changes to the timing and duration of rains. A farmer
from Rupa Sub-county, Natopojo village (Farmer ID no. 2) explains,
“We used to have specific months during which we had to prepare gardens,
dig, weed, chase birds, and harvest. Today, the weather is unpredictable; we
never know what to expect. This makes it hard to plan for our gardens. We
139
do not have a clear planting calendar and are not sure of what the weather
condition will be like tomorrow. We are also noticing signs of climatic
change. For example, we now experience a 6-month period of drought,
followed by 6 months of rainfall, which was not the case before.”
When farmers were asked to recall how many droughts had occurred within set
timeframes, 52% indicated they had experienced 1 to 3 droughts, 38% (8) indicated
that 2000 and 2014 had been drought years, and 61% said that they had experienced 1
to 3 droughts between 2000 and 2009. About 38% recalled 1 to 3 droughts between
1990 and 2000, and only 23% indicated 1 to 3 droughts between 1980 and 1990. The
data can be interpreted at least in two ways: 1) from the farmers’ perspective droughts
have become more frequent, 2) farmers tend to recall more recent droughts since the
number of droughts recalled increase with time (Figure 5.7).
It is evident from the farmers’ perspective that droughts have become more
frequent and the negative impacts are increasing. For example, one farmer mentioned
that droughts would probably get worse due to extensive deforestation that is
negatively affecting the climate. Another farmer stated that extractive coping
strategies, including mining and charcoal burning, are increasing the risk of drought.
In the words of a farmer from Katikekile (Farmer ID. 6),
“There is no future for the next generations because when drought occurs,
people resort to deforestation, gold mining, and hunting wild animals, all of
which leads to the destruction of the environment. The future is doomed by
these kinds of conditions. People die due to hunger and very few people can
140
1980-1990 1990-2000 2000-2010 2010-2014
0
5
10
15
20
25
30
35
40
45
50
55
60
65
Period
P
er
ce
nt
of
F
ar
m
er
s
Don’t Know 1 to 3 droughts 4 to 5 droughts
6 to 10 droughts average no. of droughts recalled*10
Figure 5.7: Farmer recollection of droughts.
survive such conditions.”
5.4.4 Coping with drought in Moroto
Based on interactions with people with deep knowledge of Karamoja, it is clear
that droughts, mild or severe, have serious impacts on communities in Moroto. The
quotes below include excerpts from some of the interviews summarizing some of the
impacts of drought as perceived by farmers in Moroto.
“People are malnourished and weak due to lack of food”.... “Pastoralists
lack strength to move their animals to search of pasture and water”.....
“Many people resort to eating animal skins”. “Outbreak of human diseases
such as diarrhea and malaria, cholera, diarrhea, yellow fever, and river
141
blindness increase death rates...”. “Death rates are high among the elderly
and young children.
The people of Karamoja cope with impacts of drought and other natural
disasters in a variety of ways to sustain their livelihoods. From our discussions with
farmers, it appears that during periods of drought people exhaust all possible means
available to them for survival. Many abandon their fields to look for alternative
sources of livelihoods. People in Karamoja are increasingly involved in casual work,
including stone quarrying and gold mining. Many women sell local beer for income;
some engage in tree cutting, the sale of livestock and charcoal, the gathering of wild
foods, and cattle raids. Due to lack of food, families reduce the number of meals
consumed per day. Typically, families have three meals a day, but during a famine,
meals are reduced to one for the strong and two for young children and the elderly.
The reduced food intake and lack of access to nutritious food results in outbreaks of
disease such as diarrhea, cholera, and trachoma. Many eat food residue such as banana
and cassava peels (Figure 5.7), while others starve, further exacerbating the spread of
disease. Shortages of food and resources also result in increased migration as people
leave in search of casual labor and other sources of income, such as pasture for
livestock, markets for livestock and charcoal, and areas to cut trees and gather wild
foods. The alternative to these coping strategies to sustain livelihoods is to rely on
food aid and other assistance from the government and development partners. These
at best are effective only in the short-term and are very destructive in the long run as
they build dependency.
142
Figure 5.8: Left: A lady reveals the only food she has in storage - dry leaves from a
local tree. Top left: Sun-drying cassava peels. Bottom right: villagers from
Kadilakeny carrying rocks to cover up erosion gullies as part of a ‘Food for
Work Program’ in January 2016.
143
5.5 Discussion
Farmer’s perceptions of drought in Moroto are considerably conflated with other
factors that are not climate-related. These include armed conflict, the protracted
disarmament campaign, intensity of cattle raids, and additional livelihood impacts
such as sales of other assets, indebtedness, out-migration, and dependency on food
relief. These non-climate-related disruptive factors affect household well-being and also
contribute to crop failures, livestock deaths, and eventual famines that overshadow
farmers’ memories and experiences of the actual periods of drought. For example, due
to extreme conditions of conflict and poverty, farmers rarely recall drought years
accurately, since droughts are considered normal in comparison. Insecurity is a
substantial factor that strongly influenced farmers’ memories. Currently, it might not
be possible to disentangle drought perceptions from memories of extreme violence and
social strife. However, given that armed conflict and cattle raids are under relative
control since the late 2000’s, future studies might address this issue. It is also worth
noting that when there is little difference between drought years and non-drought years
in terms of crop production and well-being, drought years are less likely to stand out.
In addition, in the Karamojong language, ’Akoro’ drought and hunger ’Akoro’ are
homonyms.
Changes in agricultural practices in response to perceived changes have been
documented broadly for other communities, including changing planting dates,
moisture conservation, and crop selection (Kassie et al., 2013; Slegers, 2008;
Meze-hausken, 2004). Although these changes in agricultural practices may be
144
insufficient to mitigate future drought impacts, they can be enhanced by further
research and investment in promoting good practices and thus offer context for more
relevant policy making citepKassie2013. However, except for attempts to replant after
the failure of rains (for those who can afford to or have access to seeds) and income
diversification (mining and casual labor), no clear long-term strategies for managing
possible future droughts were documented in the interviews. Interviewed farmers felt
they had no control over their future. Some relied on food aid while those who owned
livestock mentioned selling it as a fallback option. This highlights a disparity that
could be exploited to support more farmers. Careful livestock restocking could widen
the safety net livestock currently provide. Government policy should aim to educate
farmers through farmer schools, on the likelihood of risks associated with extreme
drought events and continued climatic changes and encourage locally devised risk
management options. Farmer schools could have a focus on understanding and
managing climatic risks to re-engage farmers in decision-making processes leading to
better adaptation methods that are locally developed and owned.
Farmers indicated that government responses often do not meet their
expectations. Evidence also suggests that government support and response programs
fall short of what communities anticipate. Whereby not all households receive or
qualify for improved crop varieties, education programs, emergency relief and notice of
early warning signs. Other studies, including citepKassie2013, show that this
inaccessibility to affordable technologies constrains the development of promising
adaptation measures.
Information on alternative integrated approaches to reduce human suffering and
145
save lives is much needed. Efforts to induce development have failed in the past, in
large part due to the failure to recognize the complexity of the Karamojong culture
and society. Instead, most interventions have exacerbated societal problems, which
have led to conflict and continue to undermine the resilience of pastoralism
(Nakalembe et al., 2017; Mwanga, 2014; Houdet et al., 2014). There is a great need to
facilitate knowledge partnerships between the agents of intervention and the intended
beneficiaries, constructed on local peoples’ abilities achievements (Mortimore, 2005).
This could be the missing link in development planning in Karamoja.
5.6 Conclusion and policy recommendations
Aside from a vicious cycle of drought leading to resource scarcity, the Moroto
case study presents a unique geographic and sociocultural setting that includes the
protracted armed conflict, which ’necessitated’ militarized external intervention and
continuous aid. It appears from this research that the farmers have no understanding
of, nor the skills or knowledge to manage agricultural drought, having previously as
pastoralists depended mainly on livestock rather than crops (Nakalembe et al., 2017).
Not only do they lack the knowledge to manage, plan or take precautionary action
against the impacts of future droughts, but they also lack the fundamental resources
needed to make any adjustments. They are often, if not always, entirely dependent on
aid. In other words, they are not empowered to decide when, what, and how much
land to cultivate, as demonstrated in other studies of similar communities such as
those conducted by Mubaya et al. (2012); Kassie et al. (2013) and citeSlegers2008.
146
This needs to be further explored in future studies of perception to understand the
problems and costs of mal-adaptations and inaction better.
Though this study was limited in scope, it enabled identification of some
recurring themes. The affected communities are often criticized and blamed for failed
policies. However, policies fail when local conditions are poorly understood and when
the perspectives of the intended beneficiaries are not taken into account. While some
of the results confirm what is often reported in food security assessments, including
coping strategies, the interviews highlighted further nuanced the predicament of farmer
households during drought periods. Also considering the extreme poverty, the findings
of this study present interesting local perspectives on protecting the environment. It is
often the view that all communities engage in tree cutting and charcoal burning with
no concern for the environment. To the contrary, our interviews showed considerable
attention to and awareness of the destructive nature of such enterprises and their
undesirability. A deeper investigation of perceptions is warranted to highlight the
viewpoint of the affected communities. The perspectives of these communities need to
be integral to the reporting of drought impacts in the region. Perceptions of drought
are distorted due to the extreme circumstances in which the majority of farmer
households find themselves. They lack both the alternative assets to fall back on, and
the means to prepare for future events.
Studying human perceptions provides valuable perspectives of how farmer
communities cope with and mitigate the impacts of drought and climate vulnerability
(Kassie et al., 2013; Slegers, 2008; Meze-hausken, 2004). However, these efforts remain
limited and constrained by poor access to the necessary technologies and insufficient
147
support from the government. Existing opportunities to reduce production losses and
human suffering, including early warning through radio and extension services, mostly
remain ineffective yet could provide farmers valuable assistance (Kassie et al., 2013).
Future studies need to pay attention to how different farmer communities cope with
drought. Historical, social, economic and political events that have a substantial
impact on perceptions need to be considered. A better understanding is needed of how
such events, such as internal conflicts resulting in mass migrations, and policies
affecting land use practices; inevitably leave an impact on perceptions and influence
future responses.
This study emphasizes the need to improve communication and linkages between
development planners and the local communities. There is a need to facilitate
knowledge partnerships, constructed on local people’s achievements, between the
agents of intervention and the intended beneficiaries (Mortimore, 2005). Filling the
knowledge gap between insiders and outsiders is very important. Outsiders who are
more aware of the harsh realities of managing relatively unproductive natural resources
at high levels of risk can, therefore, facilitate better interventions that are
evidence-based. Adaptation measures should be based on locally developed options,
and where these are insufficient or non-existent efforts to re-empower these
communities need to take priority. Farmers should not passively adopt externally
developed agendas and should always be fully engaged in the processes, policies, and
decisions that affect their communities. Any measures that run counter to the cultural
values of the community will likely fail in the long-run and will continue to limit
adoption and development of better drought management practices.
148
This work also illustrates that perceptions of drought by un-empowered farmers
that experience one extreme after another are unique and pose greater challenges for
policy making. A situation, where farmers are unaware of the risks they face and
assume drought management is solely the government’s responsibility, is limited in
terms of community response options to drought and only yields a no-win situation.
As illustrated in this study, there is a disconnect between perceptions and actions/
decisions taken. Because the majority are extremely poor, only limiting or reducing
access to food aid will not likely yield the desired outcomes of encouraging
independence. Unique sets of programs are needed to support and help farmers develop
their adaptation strategies supplemented with appropriate technologies. Improvements
to farmer’s livelihoods through local crop production need to be demonstrated to
dispel the fatalistic perception of helplessness. Given the unreliable nature of rainfall
and climate projections of worsening conditions, rain-fed agriculture appears to be
unviable. A program of carefully managed groundwater irrigation combined with
extensive use of drought resistant varieties may be the only viable solution for this
region. Such a program would need considerable investment and would need to be
implemented in a way that would fully engage the community and educate farmers to
avoid the pitfalls of salinization. This study also further demonstrates the need to use
both physical and social science data when designing policies. Historical records and
projected changes can be used to educate farmers on risk management, while social
science can highlight opportunities for behavioral change.
149

6 Conclusion
6.1 Summary
The five individual studies in this dissertation integrate remote-sensing, survey,
field, interview data and reports to address the research objectives presented in Section
1.1 of this dissertation. The overall goal of this research was to better understand
vulnerability and to quantitatively assess drought and its impacts on agriculture and
the people of Karamoja in northeastern Uganda.
Chapter 2 of this dissertation titled ”Mapping Drought Vulnerability in Uganda”
applied a multi-scale, spatially explicit, quantitative approach to drought vulnerability
assessment within Uganda. The three dimensions of vulnerability: exposure, sensitivity
and adaptive capacity; were analyzed for Uganda’s 11 sub-regions using rainfall
(station and satellite derived), remote-sensing and socio-economic data from the most
recent Uganda National Panel Survey (UNPS) of 2014 available from the World Bank
Living Standards Measurement Study (LSMS) Website. The methods presented in this
chapter are scalable and easily reproducible, highlighting opportunities for strategic
interventions. Although the Karamoja Region is moderately exposed to drought, it is
151
highly sensitive and by far has the least adaptive capacity in Uganda. The region is
therefore the most vulnerable to agricultural drought. Adaptive capacity is the main
contributor to variation in drought vulnerability between regions. While Karamoja has
experienced more droughts than any other region in the country in the last 30 years,
other regions including Teso, Lango and West Nile have experienced severe drought
more frequently. In terms of sensitivity to drought, Karamoja and Teso are extremely
sensitive to drought, implying that meteorological droughts often have severe impacts
on crops/vegetation in the region as was indicated by the satellite-derived NDVI data.
Moreover, Karamoja has the lowest drought adaptive capacity, having the highest
poverty rates and a higher dependency on the natural environment for livelihood. In
totality, the Karamoja region is at least twice more vulnerable to drought than any
other region in Uganda. The results indicate that unless strong programs are put in
place to increase the adaptive capacity (particularly, poverty alleviation in Northern
Uganda), the impact of severe drought will continue to prevent development of
sustainable livelihoods. This study also highlights the need for a more continued
monitoring of drought at the sub-regional level in Karamoja to better understand its
temporal and spatial severity across the region where more and more agro-pastoralist
are reverting to pure crop-based livelihoods aided by policy and programming.
Chapter 3 titled ”Agricultural Land Use Change in Karamoja Region, Uganda”
examined cropland expansion in Karamoja, the region most vulnerable to drought in
Uganda. Dramatic cropland expansion is examined by investigating the links between
biophysical and political historical events leading to the current state of agricultural
land use. The objective was to quantify agricultural expansion, uncover the dominant
152
drivers leading to the current state of agricultural land-use and its impacts on
livelihoods. Region-wide analysis of remotely-sensed data, land-use policy and history
as well as farmer interviews were undertaken. Results indicate that government
programs instituting sedentary agriculture are the most significant drivers of cropland
expansion in Karamoja. We show a 299% increase in cropland area between 2000 and
2011 with most expansion occurring in Moroto District (from 706 hectares to 23,328
hectares). The study found no evidence of an increase in overall crop production or
food security and food aid continues to be essential due to recurrent crop failures. Due
to lack of resources for inputs (e.g., seeds and labour) cultivated fields remain very
small in size and over 55% of once cultivated land is left fallow. The findings bring into
question whether continued promotion of rain-fed agriculture in Karamoja serves the
best interests of the people. Current cropland expansion is directly competing with
and compromising pasture areas critical for livestock-based livelihoods. Without strong
agricultural extension programs and major investments in climate-smart options suited
to drought, cropland expansion will continue to have a net negative impact, especially
in the context of current climate projections which indicate a future decrease in rainfall,
increase in temperature and an increase in the frequency and magnitude of extreme
events. The chapter presents a detailed historical account of the factors leading to the
current state of land use. Using very high resolution satellite data, an agricultural land
use map for the region was generated. The derived map provided concrete proof of the
unprecedented rate of cropland expansion. This expansion will likely continue
irrespective of the fact that productivity remains negligible. Underlying forces (e.g.
cropland expansion programs and controlled grazing) originating from land use policy
153
and development programs remain the major drivers of this expansion, rather than
proximate causes (direct local level actions). From this study it was evident that the
process, patterns and impacts of land use change remain of secondary importance to
development planners in the region, and if current programs are continued, agricultural
expansion will continue, although the future rates of change and extent are unknown.
Karamoja is notoriously food-insecure and has been the recipient of food aid for
most years during the last two decades. One of the main factors causing food
insecurity is drought. Reliable, area-wide, long-term data for detecting and monitoring
drought conditions are critical for timely, life-saving interventions and the long-term
development of the region, yet such data are sparse or unavailable. Due to advances in
satellite remote sensing, characterizing drought in data sparse regions like Karamoja
has become possible. Chapter 4 titled ”Characterizing Agricultural Drought in the
Karamoja sub-region of Uganda with Meteorological and Satellite-Based Indices”
enabled a comprehensive understanding of agricultural drought and its impact on the
people of Karamoja using proxy data. In this study, agricultural droughts are
characterized and the suitability of NDVI-based drought monitoring was evaluated.
The study showed that in comparison to the existing data, NDVI data currently
provide the best, consistent and spatially-explicit information for operational drought
monitoring in Karamoja. Results indicate that the most extreme agricultural drought
in recent years occurred in 2009 followed by 2004 and 2002. Results also suggest that
in Karamoja, moderate to severe droughts (e.g. 2008) often have the same impact on
crops and human needs (e.g. food aid) as extreme droughts (e.g. 2009). As a
proof-of-concept in this chapter, I explored a quantitative method to estimate the
154
number of people needing food aid and the population likely to fall under the
Integrated Food Security Phase Classification (IPC) Phase 3 (crisis) due to drought
severity. The model indicates that 90.7% of the variation in the number of people
needing aid can be explained by NDVI data and that approximately 759,000 people
will require food aid to sustain their consumption until the 2017 harvest in August.
This simple method has the potential to inform policy and support government
planning and decision-making. The biggest drivers of food insecurity are; the
cultivation of crops on marginal land with insignificant inputs, the lack of irrigation,
and previous systematic incapacitation of livestock (pastoral) alternatives through
government programing. Further research is needed to bridge empirical results with
social economic studies on drought impacts on communities in the region to better
understand additional factors e.g. cultural values and gender roles that will need to be
addressed to ensure livelihood resilience. This research was enabled by the map data
derived in Chapter 3 that was used to mask non-agricultural areas when analyzing
satellite derived NDVI data. The majority of droughts between 1980 and 2012 were
mild or moderate with the exception of 2008, when the region experienced a severe
drought and 2009 when there was an extreme drought. This part of the research
underscored the value of using quantitative data in drought assessments. Quantitative
data, particularly satellite data are also more reliable and timely in providing early
warning and planning responses even in Karamoja. The work in this chapter, enabled
me to take an active role in designing the Disaster Risk Financing program that was
launched in February 2017 by the Office of the Prime Minister.
Chapter 5 of this dissertation is a pilot study entitled ”Human Perceptions of
155
Drought in Karamoja Uganda: a Moroto District Case Study” focused on
understanding farmers’ behavioral responses to drought and its impacts in the region.
Studying the human perception of drought is one way of understanding the behavioral
responses of communities affected by droughts (Dagel, 2010; Taylor et al., 1988). Such
responses are ideally manifested in adaptation and mitigation measures.
Understanding human perceptions of drought can facilitate better interventions that
are evidence-based by outsiders, who are unaware of the harsh realities of managing
relatively unproductive natural resources at high levels of risk. Although both drought
exposure and sensitivity are high, to the best of my knowledge there have been no
studies on human perceptions of drought in Karamoja Region. This chapter presents
the results of a study of human perceptions of drought in Moroto District in Karamoja
sub-region. Following the four critical elements of drought perception, this chapter
combines the results from interviews and observations from remotely-sensed
information. Results indicate that past experiences of drought do not necessarily
impact future expectations of drought in the district. As anticipated, more experienced
farmers tend to consider more climatologically severe droughts as drought years
compared to less experienced young framers. From the interviews it appeared that no
real long-term adjustment plans have been made by the farmers. From time to time
the majority of them depend on emergency food assistance. The results indicated that
factors such as; conflict (insecurity) and interventions by government and its partners
intermingle with culture to have profound direct influence on farmers’ perception of
drought amongst communities in Moroto district. Many of the farmers interviewed
considered food aid as a coping strategy, only second to livestock. External to farmers
156
practices, these factors continue to disrupt community practices, prohibiting adoption
or evolution of better management practices. Interview information was related to
drought information obtained from satellite derived information for the district. In
doing so I discovered that often food security reports under-represent the actual
circumstance and the predicament of affected communities. During drought periods
many households are despondent but this is rarely captured in time for proper
intervention. This work is a departure from commonly held beliefs in Uganda about
the people of Karamoja, that consider the behavior irrational and destructive. This
study has proved that communities are aware and concerned and nullified claims that
the people of Moroto without any care or awareness of the consequences, are
destroying the very natural environment they depend on.
6.2 Future Research
Although Karamoja remains a priority area for the Ugandan Government and for
many international aid agencies and organizations, they are still failing to achieve most
of their project/ program goals. As highlighted in Chapter 3 and 4, there is great need
for more empirical research on Karamoja. There are huge gaps in knowledge on several
issues such as sustainable pastoralism, the proliferation of temporary labor
opportunities through programs such as food for work and in the long-term breakdown
of the traditional community fabric due to shifting gender roles. This will likely
continue to hinder development in the region.
One of the directions for future research is to develop a dense time-series on
157
landcover and land use changes. In this research the landuse map was derived using
very high-resolution data that are not freely accessible and have limited coverage.
Updated landcover information can better inform planning and further improve
understanding of the impacts of drought in the region. Future research should focus on
deriving land use information from newer satellite systems such as Sentinel 2. Sentinel
2 data are freely available at 10m resolution1. At this resolution and considering the
system revisit time of 5 days, it is feasible to derive an annual cropland mask, which
would be helpful for regional planning and food security monitoring.
Research conducted as part of this dissertation and the methods presented can
help inform the development of a national drought monitoring system, since all data
used are freely accessible. However, there is still considerable room to improve the
basis of drought monitoring in Karamoja and for Uganda as a whole. For example,
there are several model datasets available for the broader East Africa such as those
produced by FEWSNET. Future research should focus on sub-regional validation and
calibration of these datasets as well as transitioning these data and tools into
operational use by government agencies.
Chapter 5 of this dissertation shows that there is still much to learn about
Karamoja and using a mixed methods approach provides a broader perspective of this
complex landscape. Future research can expand research on perceptions of drought to
cover the entire region and obtain data from a more representative sample. Expanding
the scope (sample size and geographic) to cover other district in the region would
1See Sentinel 2 Users Guide avaible at: https://earth.esa.int/web/sentinel/user-guides/sentinel-2-
msi/resolutions/spatial
158
better capture the diversity in perceptions of drought in the region. Understating
perception in the region offers insight into the range of judgments, beliefs, and
attitudes that underpin human decisions and choices which lead to actions that affect
the environment and influence landscape scale changes, that inevitably cascade into
further decisions and impacts (Taylor et al., 1988).
6.3 Policy Implications and Next Steps
If current stability (in terms of security) is maintained, it is very likely that
Karamoja’s population will continue to grow. Population and human activity,
particularly near towns, will continue to increase. Moroto town for example, has
become more urbanized due to an increase in demand for services by the employees of
various local and international organizations. The question that remains is; how will
the growing food demand be met? Market prices for food items are already far beyond
the means of the local communities. Short-term strategies should aim to diversify
income sources for communities to inevitably reduce dependence on rain-fed
crop-based livelihood. The interviews revealed that farmers with fall back options,
particularly livestock, tended to do much better during drought years. There is
therefore, an opportunity to invest in livestock-based livelihoods that are more suited
to large part of the region.
The results strongly indicate that unless well thought through programs are put
in place to increase adaptive capacity, particularly focusing on eradicating poverty in
northern Uganda through alternative livelihood streams, severe drought impacts will
159
continue to severely affect communities. Therefore, there is greater need for increased
attention toward improving drought management at national and sub-national level. A
key first step would include compilation of comprehensive and reliable information on
drought impacts. This information would be essential for improved understanding of
Uganda’s vulnerability to drought. It would also justify the need for increased
investment in less drought sensitive sectors or devise mitigation strategies. Without
this information, it will remain difficult to convince policy and other decision-makers of
the need for additional investments in drought monitoring and prediction, mitigation,
and preparedness.
Although cultivated area has dramatically increased, we find no quantifiable
overall increase in yield, or per-capita production evidenced by the consistent food
insecurity. This status quo, (poor yields and dependence on food aid) is likely to be
maintained as more land is put to crop cultivation by poor households and meager
investments are made in livestock-based livelihood opportunities. From the findings in
this dissertation, it appears that given the climate in the region and the nature of soils,
water for irrigation has to be secured for crop cultivation to have an impact on food
security. This requires a lot of financial investment due to the high costs involved and
low water availability in Karamoja and would require careful land management
(Avery, 2014). Finding a balance between crop-based (where suitable) with
livestock-based livelihood should be of primary importance in development planning
for Karamoja. This will require empirical understanding of economic, environmental
and ideological differences which should guide development planning and investments.
Since 2012, I have had multiple opportunities to engage with multiple
160
stakeholders within Karamoja (mostly in Moroto, Napak and Kotido Districts) and at
the national level. These opportunities have enabled me to better understand the
region and explore my research questions. My observation is that many outsiders do
not understand the predicament of the communities they are working with in
Karamoja. As a result the region’s problems are often assessed on face value and quick
fixes that are often prescribed and are not sustainable. It is remarkable, that the
majority of the people of Karamoja have and continue to sustain their communities in
an environment of policy-failure and urban bias. At one level, this speaks to the
resilience of the Karamojong. There is need for strategic, informed and native-driven
policy to counter current problems. Pastoralists’ knowledge and achievements should
be central because they posses accumulated experience of their specific dry land
environment (Mortimore, 2005, 1989). Ownership of any strategy could take pastoral
communities further than any interventions. The fact that mobility is important to
pastoral Karamojong, approaches that can be beneficial to them could include
streamlining property rights in favor of flexibility for controlled movement based on
real-time monitoring of pasture conditions. The Uganda government needs to develop
policies that promote balanced interventions. More urgently alternatives need to be
developed to discourage crippling response measures such as permanent supply of food
aid, to foster and enable adaptation.
There is a need to facilitate knowledge partnerships, constructed on local
people’s achievements between the agents of intervention and the intended beneficiaries
(Mortimore, 2005). Filling the knowledge gap between insiders and outsiders is very
important. Careful observation and analysis of Karimojong economic behavior
161
indicates a high degree of rationality both for individual actors and for the economic
system (Quam, 1978). Outsiders who are aware of the harsh realities of managing
relatively unproductive natural resources at high levels of risk can therefore facilitate
better interventions that are evidence-based. A better solution would be to establish
adaptation measures based on locally developed options with the people of Karamoja
fully engaged in the processes and policies that affect their communities. Any measures
that defy the cultural values are destined to fail.
162
163
Appendix
Appendix A: Institutional Review Board Approval
Document 363568-3
 
- 1 - Generated on IRBNet
1204 Marie Mount Hall
College Park, MD 20742-5125
TEL 301.405.4212
FAX 301.314.1475
irb@umd.edu
 
www.umresearch.umd.edu/IRB
 INSTITUTIONAL REVIEW BOARD
 
DATE: March 13, 2014
  
TO: Catherine Nakalembe, PHD
FROM: University of Maryland College Park (UMCP) IRB
  
PROJECT TITLE: [363568-3] AGRICULTURAL LAND USE, DROUGHT IMPACTS AND
MITIGATION: A REGIONAL CASE STUDY FOR KARAMOJA, UGANDA
REFERENCE #:  
SUBMISSION TYPE: Continuing Review/Progress Report
  
ACTION: APPROVED
APPROVAL DATE: March 13, 2014
EXPIRATION DATE: June 24, 2015
REVIEW TYPE: Expedited Review
  
REVIEW CATEGORY: Expedited review category # 7
  
Thank you for your submission of Continuing Review/Progress Report materials for this project. The
University of Maryland College Park (UMCP) IRB has APPROVED your submission. This approval is
based on an appropriate risk/benefit ratio and a project design wherein the risks have been minimized. All
research must be conducted in accordance with this approved submission.
This submission has received Expedited Review based on the applicable federal regulation.
Please remember that informed consent is a process beginning with a description of the project and
insurance of participant understanding followed by a signed consent form. Informed consent must
continue throughout the project via a dialogue between the researcher and research participant. Federal
regulations require each participant receive a copy of the signed consent document.
Please note that any revision to previously approved materials must be approved by this committee prior
to initiation. Please use the appropriate revision forms for this procedure which are found on the IRBNet
Forms and Templates Page.
All UNANTICIPATED PROBLEMS involving risks to subjects or others (UPIRSOs) and SERIOUS and
UNEXPECTED adverse events must be reported promptly to this office. Please use the appropriate
reporting forms for this procedure. All FDA and sponsor reporting requirements should also be followed.
All NON-COMPLIANCE issues or COMPLAINTS regarding this project must be reported promptly to this
office.
This project has been determined to be a Minimal Risk project. Based on the risks, this project requires
continuing review by this committee on an annual basis. Please use the appropriate forms for this
procedure. Your documentation for continuing review must be received with sufficient time for review and
continued approval before the expiration date of June 24, 2015.
Please note that all research records must be retained for a minimum of three years after the completion
of the project.
164
Appendix B: Institutional Review Board Approval
Document 363568-2
 
- 1 - Generated on IRBNet
1204 Marie Mount Hall
College Park, MD 20742-5125
TEL 301.405.4212
FAX 301.314.1475
irb@umd.edu
 
www.umresearch.umd.edu/IRB
 INSTITUTIONAL REVIEW BOARD
 
DATE: June 25, 2013
  
TO: Catherine Nakalembe, PHD
FROM: University of Maryland College Park (UMCP) IRB
  
PROJECT TITLE: [363568-2] AGRICULTURAL LAND USE, DROUGHT IMPACTS AND
MITIGATION: A REGIONAL CASE STUDY FOR KARAMOJA, UGANDA
REFERENCE #:  
SUBMISSION TYPE: Revision
  
ACTION: APPROVED
APPROVAL DATE: June 25, 2013
EXPIRATION DATE: June 24, 2014
REVIEW TYPE: Expedited Review
  
REVIEW CATEGORY: Expedited review category # 6 & 7
  
Thank you for your submission of Revision materials for this project. The University of Maryland College
Park (UMCP) IRB has APPROVED your submission. This approval is based on an appropriate risk/
benefit ratio and a project design wherein the risks have been minimized. All research must be conducted
in accordance with this approved submission.
This submission has received Expedited Review based on the applicable federal regulation.
Please remember that informed consent is a process beginning with a description of the project and
insurance of participant understanding followed by a signed consent form. Informed consent must
continue throughout the project via a dialogue between the researcher and research participant. Federal
regulations require each participant receive a copy of the signed consent document.
Please note that any revision to previously approved materials must be approved by this committee prior
to initiation. Please use the appropriate revision forms for this procedure which are found on the IRBNet
Forms and Templates Page.
All UNANTICIPATED PROBLEMS involving risks to subjects or others (UPIRSOs) and SERIOUS and
UNEXPECTED adverse events must be reported promptly to this office. Please use the appropriate
reporting forms for this procedure. All FDA and sponsor reporting requirements should also be followed.
All NON-COMPLIANCE issues or COMPLAINTS regarding this project must be reported promptly to this
office.
This project has been determined to be a Minimal Risk project. Based on the risks, this project requires
continuing review by this committee on an annual basis. Please use the appropriate forms for this
procedure. Your documentation for continuing review must be received with sufficient time for review and
continued approval before the expiration date of June 24, 2014.
Please note that all research records must be retained for a minimum of three years after the completion
of the project.
165
Appendix C: Uganda National Council for Science and
Technology Research Approval Ref: 473
166
Bibliography
W. Adger, S. Agrawala, M. Mirza, C. Conde, K. O’Brien, J. Pulhin, R. Pulwarty,
B. Smit, and K. Takahashi. Summary for Policymakers. In Intergovernmental Panel
on Climate Change, editor, Climate Change 2013 - The Physical Science Basis,
pages 1–30. Cambridge University Press, Cambridge, 2007. ISBN 978 0521 88010-7.
doi: 10.1017/CBO9781107415324.004.
W. N. Adger. Vulnerability. Global Environmental Change, 16(3):268–281, 2006. ISSN
09593780. doi: 10.1016/j.gloenvcha.2006.02.006.
A. AghaKouchak, A. Farahmand, F. S. Melton, J. Teixeira, M. C. Anderson, B. D.
Wardlow, and C. R. Hain. Remote sensing of drought: Progress, challenges and
opportunities. Reviews of Geophysics, 53(3):1–29, 2015. ISSN 87551209. doi:
10.1002/2014RG000456.Received.
I. Alcantara-Ayala. Geomorphology , Natural Hazards , Vulnerability and Prevention
of Natural Disasters in Developing Countries. Geomorphology, 47(September 2016):
107–124, 2002. ISSN 0169555X. doi: 10.1016/S0169-555X(02)00083-1.
W. Alley. The Palmer drought severity index: limitations and assumptions. Journal of
Climate and Applied Meteorology, 23:1100–1109, 1984.
E. Allison, A. L. Perry, M. Badjeck, W. N. Adger, K. Brown, D. Conway, A. S. Halls,
G. M. Pilling, J. D. Reynolds, N. L. Andrew, and N. K. Dulcy. Vulnerability of
national economics to the impacts of climate change on fisheries. Fish and Fisheries,
10:173–196, 2009.
P. Antwi-Agyei, E. D. Fraser, A. J. Dougill, L. C. Stringer, and E. Simelton. Mapping
the vulnerability of crop production to drought in Ghana using rainfall, yield and
socioeconomic data. Applied Geography, 32(2):324–334, mar 2012. ISSN 01436228.
doi: 10.1016/j.apgeog.2011.06.010.
A. Anyamba and C. Tucker. From El Nino to La Nina: vegetation response patterns
over East and Southern Africa during the 1997-2000 period. Journal of Climate,
pages 3096–3103, 2002.
A. Anyamba, C. Tucker, and J. Eastman. NDVI anomaly patterns over Africa during
the 1997/98 ENSO warm event. International Journal of Remote Sensing, 22(10):
1847–1859, jan 2001. ISSN 0143-1161. doi: 10.1080/01431160010029156.
167
C. Atzberger. Advances in remote sensing of agriculture: Context description, existing
operational monitoring systems and major information needs. Remote Sensing, 5(2):
949–981, 2013. ISSN 20724292. doi: 10.3390/rs5020949.
S. Avery. Water Development and Irrigation in Karamoja , Uganda. Review of Water
Development and Irrigation in Karamoja, Uganda. Technical Report February,
DanChurchAid, 2014.
S. Ayoo, R. Opio, and O. T. Kakisa. Karamoja Situational Analysis. Technical Report
January, Care International In Uganda, 2013.
S. Baker. A Background to the Study of Drought in East Africa. African Affairs, 73
(291):170–177, 1974.
S. Barrios, B. Ouattara, and E. Strobl. The impact of climatic change on agricultural
production: Is it different for Africa? Food Policy, 33(4):287–298, aug 2008. ISSN
03069192. doi: 10.1016/j.foodpol.2008.01.003.
I. Becker-Reshef, C. Justice, M. Sullivan, E. Vermote, C. Tucker, A. Anyamba,
J. Small, E. Pak, E. Masuoka, J. Schmaltz, M. Hansen, K. Pittman, C. Birkett,
D. Williams, C. Reynolds, and B. Doorn. Monitoring Global Croplands with Coarse
Resolution Earth Observations: The Global Agriculture Monitoring (GLAM)
Project. Remote Sensing, 2(6):1589–1609, jun 2010a. ISSN 2072-4292. doi:
10.3390/rs2061589.
I. Becker-Reshef, E. Vermote, M. Lindeman, and C. Justice. A generalized
regression-based model for forecasting winter wheat yields in Kansas and Ukraine
using MODIS data. Remote Sensing of Environment, 114(6):1312–1323, jun 2010b.
ISSN 00344257. doi: 10.1016/j.rse.2010.01.010.
J. Bevan. Crisis in Karamoja, 2008.
R. Blench. You can’t go home again’ Pastoralism in the new millennium. Technical
Report 17, Overseas Development Institute, London, UK, 2001. URL
http://www.cispal.fsnet.co.uk.
P. Boateng. Land Access, Agricultural Land Use Changes and Narratives about Land
Degradation in the Savannahs of Northeast Ghana during the Pre-Colonial and
Colonial Periods. Social Sciences, 6(1):35, 2017. ISSN 2076-0760. doi:
10.3390/socsci6010035. URL http://www.mdpi.com/2076-0760/6/1/35.
N. Brooks, W. N. Adger, and P. M. Kelly. The determinants of vulnerability and
adaptive capacity at the national level and the implications for adaptation. Global
Environmental Change, 15(2):151–163, 2005. ISSN 09593780. doi:
10.1016/j.gloenvcha.2004.12.006.
M. E. Brown and K. DeBeurs. Evaluation of multi-sensor semi-arid crop season
parameters based on NDVI and rainfall. Remote Sensing of Environment, 112(5):
2261–2271, may 2008. ISSN 00344257. doi: 10.1016/j.rse.2007.10.008.
168
T. Bunting and L. Guelke. Behavioral and Perception Geography: A Critical
Appraisal. Annals of the Association of American Geographers, 69(3):448–462, 1979.
J. Burns, G. Bekele, and D. Akabwai. Livelihood Dynamics in Northern Karamoja.
Technical report, Feinstein International Center, 2013.
R. Cervigni and M. Morris. Confronting Drought in Africa’s Drylands: Opportunities
for Enhancing Resilience. Openknowledge.Worldbank.Org, page 221, 2016. ISSN
1098-6596. doi: 10.1596/978-1-4648-0817-3.
J. H. Christensen, B. Hewitson, A. Busuioc, A. Chen, X. Gao, I. M. Held, R. Jones,
R. K. Kolli, W.-T. Kwon, R. Laprise, V. Magana Rueda, L. O. Mearns, C. G.
Menendez, J. Raisanen, A. Rinke, A. Sarr, P. Whetton, and C. L. Authors. Regional
Climate Projections. In Climate Change 2007: The Physical Science Basis.
Contribution of Working Group I to the Fourth Assessment Report of the
Intergovernmental Panel on Climate Change, chapter 11, pages 847–940. Cambridge
University Press, Cambridge, United Kingdom, 2007. ISBN 9780521880091. doi:
10.1080/07341510601092191.
J. E. Cinner, T. R. McClanahan, N. A. J. Graham, T. M. Daw, J. Maina, S. M. Stead,
A. Wamukota, K. Brown, and O. Bodin. Vulnerability of coastal communities to key
impacts of climate change on coral reef fisheries. Global Environmental Change, 22
(1):12–20, 2012. ISSN 09593780. doi: 10.1016/j.gloenvcha.2011.09.018.
S. Coldham. Land reform and customary rights: the case of Uganda. Journal of
African Law, 44(1):65–77, 2000. ISSN 0021-8553. doi: 10.1017/S0021855300012043.
S. Cutter, L. Barnes, M. Berry, C. Burton, E. Evans, E. Tate, and J. Webb. A
place-based model for understanding community resilience to natural disasters.
Global Environmental Change, 18(4):598–606, 2008. ISSN 09593780. doi:
10.1016/j.gloenvcha.2008.07.013.
K. C. Dagel. Defining Drought in Marginal Areas: The Role of Perception. The
Professional Geographer, 49(2):192–202, 2010. ISSN 0033-0124. doi:
10.1111/0033-0124.00069.
DanChurchAid. Climate change and adaptation in the Karamoja Sub-Region.
Technical report, DanChurchAid, Kampala, Uganda, 2010.
W. Deshler. Livestock trypanosomiasis and human dettlement in northeastern Uganda.
Geographical Review, 50(4):541–554, 1960. ISSN 00167428.
D. M. Diggs. Drought Experience and Perception of Climatic Change among Great
Plains Farmers. Journal of Natural and Social Sciences, 1:114–132, 1991.
Directorate of Relief Disaster Preparedness and Refugees. The Uganda National Policy
for Disaster Preparedness and Management. Technical Report October, Office of the
Prime Minister, Kampala, 2010.
169
D. Edwards. Characteristics of 20th Century drought in the United States at multiple
time scales. Thesis, Colorado State University, 1997.
A. Egeru, R. Osaliya, L. MacOpiyo, J. Mburu, O. Wasonga, B. Barasa, M. Said,
D. Aleper, and G.-J. Majaliwa Mwanjalolo. Assessing the spatio-temporal climate
variability in semi-arid Karamoja sub-region in north-eastern Uganda. International
Journal of Environmental Studies, 71(4):490–509, 2014. ISSN 0020-7233. doi:
10.1080/00207233.2014.919729.
J. Ellis and K. a. Galvin. Climate Patterns and Land-Use Practices in the Dry Zones
of Africa. BioScience, 44(5):340–349, 1994. ISSN 00063568. doi: 10.2307/1312384.
J. Ellis and D. Swift. The Dominant Paradigm: Pastoral Degradation of Equilibria1
Ecosystems. Journal of Range Management, 41(6):450–459, 1988a. doi:
10.2307/3899515.
J. E. Ellis and D. M. Swift. Stability of African pastoral ecosystems : Alternate
paradigms and implications for development. Journal of Rangeland Management, 41
(6), 1988b.
J. English, M. Tiffen, and M. Mortimore. Land Resource Management in Machakos
District, Kenya 1930-1990, volume 5. The World Bank, feb 1994. ISBN
978-0-8213-2734-0. doi: 10.1596/0-8213-2734-8.
J. Ferreri, B. Frei, B. Ross, and C. Stoker. Pastoralists, Peace and Livelihoods:
Economic interventions to build peace in Karamoja, Uganda. Technical Report May,
The Elliot School for International Affairs, 2011.
FEWS NET. Famine Early Warning Systems Network, 2015. URL
http://www.fews.net/east-africa/uganda http://www.fews.net/.
FEWSNET. Conflict Baseline Study Report Conducted in the Karamajong Cluster of
Kenya and Uganda. Technical Report August, USAID Famine Early Warning
System Network (USAID/FEWSNET), 2005.
FEWSNET. About Us — Famine Early Warning Systems Network, 2017. URL
http://www.fews.net/about-us.
M. Fisher and S. Snapp. Smallholder Farmers’ Perceptions Of Drought Risk And
Adoption Of Modern Maize In Southern Malawi. Experimental Agriculture, 50(04):
533–548, 2014. ISSN 0014-4797. doi: 10.1017/S0014479714000027.
I. D. Flitcroft, J. R. Milford, and G. Dugdale. Relating Point to Area Average Rainfall
in Semiarid West Africa and the Implications for Rainfall Estimates Derived from
Satellite Data. Journal of Applied Meteorology, 28(4):252–266, apr 1989. ISSN
0894-8763. doi: 10.1175/1520-0450(1989)028¡0252:RPTAAR¿2.0.CO;2.
170
M. M. Fontaine and A. C. Steinemann. Assessing Vulnerability to Natural Hazards:
Impact-Based Method and Application to Drought in Washington State. Natural
Hazards Review, 10(1):11–18, 2009. ISSN 1527-6988. doi:
10.1061/(ASCE)1527-6988(2009)10:1(11).
B. Franch, E. F. Vermote, I. Becker-Reshef, M. Claverie, J. Huang, J. Zhang,
C. Justice, and J. A. Sobrino. Improving the timeliness of winter wheat production
forecast in the United States of America, Ukraine and China using MODIS data and
NCAR Growing Degree Day information. Remote Sensing of Environment, 161:
131–148, 2015. ISSN 00344257. doi: 10.1016/j.rse.2015.02.014.
E. Fratkin. East African pastoralism in transition: Maasai, Boran, and Rendille cases.
African Studies Review, 44(3):1–25, 2001. ISSN 00020206. doi: 10.2307/525591.
M. Friedl, D. McIver, J. Hodges, X. Zhang, D. Muchoney, A. Strahler, C. Woodcock,
S. Gopal, A. Schneider, A. Cooper, A. Baccini, F. Gao, and C. Schaaf. Global land
cover mapping from MODIS: algorithms and early results. Remote Sensing of
Environment, 83(1-2):287–302, nov 2002. ISSN 00344257. doi:
10.1016/S0034-4257(02)00078-0.
S. Fritz, L. See, and F. Rembold. Comparison of global and regional land cover maps
with statistical information for the agricultural domain in Africa. International
Journal of Remote Sensing, 31(9):2237–2256, may 2010. ISSN 0143-1161. doi:
10.1080/01431160902946598.
S. Fritz, L. See, I. a. N. Mccallum, L. You, A. Bun, E. Moltchanova, M. Duerauer,
F. Albrecht, C. Schill, C. Perger, P. Havlik, A. Mosnier, P. Thornton,
U. Wood-sichra, M. Herrero, and I. Becker-Reshef. Mapping global cropland and
field size. Global Change Biology, pages 1–13, 2015. doi: 10.1111/gcb.12838.
C. Funk, J. Rowland, G. Eilerts, and L. White. Famine Early Warning Systems
Network: Informing Climate Change Adaptation Series A Climate Trend Analysis of
Uganda. Technical report, Famine Early Warning Systems Network, 2012.
C. Funk, P. Peterson, M. Landsfeld, D. Pedreros, J. Verdin, S. Shukla, G. Husak,
J. Rowland, L. Harrison, A. Hoell, and J. Michaelsen. The climate hazards infrared
precipitation with stationsa new environmental record for monitoring extremes.
Scientific Data, 2:1–21, 2015. ISSN 2052-4463. doi: 10.1038/sdata.2015.66.
K. A. Galvin, R. S. Reid, R. H. Behnke, and N. T. Hobbs. Fragmentation in semi-arid
and arid landscapes: Consequences for human and natural systems. In K. A. Galvin,
R. S. Reid, R. H. B. Jr, and N. T. Hobbs, editors, Fragmentation in Semi-Arid and
Arid Landscapes: Consequences for Human and Natural Systems, pages 1–411.
Springer Netherlands, Dordrecht, 2008. ISBN 9781402049064. doi:
10.1007/978-1-4020-4906-4.
171
S. Gandure, S. Walker, and J. J. Botha. Farmers’ perceptions of adaptation to climate
change and water stress in a South African rural community. Environmental
Development, 5(1):39–53, 2013. ISSN 22114645. doi: 10.1016/j.envdev.2012.11.004.
B. Gao. NDWI-A normalized difference water index for remote sensing of vegetation
liquid water from space. Remote Sensing of Environment, 266(April 1995):257–266,
1996.
B. Gartrell. Searching for ’The Roots of Famine’ The Case of Karamoja. Review of
African Political Economy, 33(August):102–110, 1985. URL
http://www.jstor.org/stable/4005676.
K. Gelsdorf, D. Maxwell, and D. Mazurana. Livelihoods, Basic Services and Social
Protection in Northern Uganda and Karamoja. Technical Report August, The
Overseas Development Institute (ODI), London, UK, 2012.
J. Gilbert. Nomadic territories: A human rights approach to nomadic peoples’ land
rights. Human Rights Law Review, 7(4):681–716, 2007. ISSN 14617781. doi:
10.1093/hrlr/ngm030.
C. Gouveia, R. Trigo, S. Beguer´ıa, and S. Vicente-Serrano. Drought impacts on
vegetation activity in the Mediterranean region: An assessment using remote sensing
data and multi-scale drought indicators. Global and Planetary Change, jun 2016.
ISSN 09218181. doi: 10.1016/j.gloplacha.2016.06.011.
Government of Uganda. Peace, Recovery and Development Plan for Northern Uganda
(PRDP). Technical Report September 2007, Uganda of Government, sep 2007.
S. Gray, M. Sundal, B. Wiebusch, M. a. Little, P. W. Leslie, and I. L. Pike. Cattle
Raiding, Cultural Survival, and Adaptability of East African Pastoralists. Current
Anthropology, 44(December):S3–S30, 2003. ISSN 0011-3204. doi: 10.1086/377669.
S. J. Gray. A memory of loss : ecological politics, local history and the evolution of
Karimojong violence. Human Organization, 59(4):401–418, 2000. ISSN 00187259.
A. D. Gregorio and L. J. M. Jansen. Land cover classification system ( LCCS ):
classification concepts and user manual. Environment and Natural Resources
Service, GCP/RAF/287/ITA Africover - East Africa Project and Soil Resources,
Management and Conservation Service, page 157, 1998.
M. C. Hansen, P. V. Potapov, R. Moore, M. Hancher, S. a. Turubanova, A. Tyukavina,
D. Thau, S. V. Stehman, S. J. Goetz, T. R. Loveland, A. Kommareddy, A. Egorov,
L. Chini, C. O. Justice, and J. R. G. Townshend. High-resolution global maps of
21st-century forest cover change. Science (New York, N.Y.), 342(6160):850–3, nov
2013. ISSN 1095-9203. doi: 10.1126/science.1244693.
R. R. Heim. A Review of Twentieth Century Drought Indices Used in the United
States, 2002.
172
J. F. Helgeson, S. Dietz, and S. Hochrainer-Stigler. Vulnerability to weather disasters:
The choice of coping strategies in rural Uganda. Ecology and Society, 18(2), 2013.
ISSN 17083087. doi: 10.5751/ES-05390-180202.
R. J. Hijmans, S. E. Cameron, J. L. Parra, P. G. Jones, and A. Jarvis. Very high
resolution interpolated climate surfaces for global land areas. International Journal
of Climatology, 25(15):1965–1978, 2005. ISSN 08998418. doi: 10.1002/joc.1276.
E. Hisali, P. Birungi, and F. Buyinza. Adaptation to climate change in Uganda:
Evidence from micro level data. Global Environmental Change, 21(4):1245–1261,
2011. ISSN 09593780. doi: 10.1016/j.gloenvcha.2011.07.005.
B. N. Holben. Characteristics of maximum-value composite images from temporal
AVHRR data. International Journal of Remote Sensing, 7(11):1417–1434, nov 1986.
ISSN 0143-1161. doi: 10.1080/01431168608948945.
J. Houdet, H. Muloopa, C. Ochieng, S. Kutegeka, and B. Nakangu. Cost Benefit
Analysis of the Mining Sector in Karamoja , Uganda. Technical report, IUCN,
Kampala, Uganda, 2014.
P. Howard, T. Davenport, and F. Kigenyi. Planning conservation areas in Uganda’s
natural forests. Oryx, 31(4):253, 1997. ISSN 0030-6053. doi:
10.1017/S0030605300022213.
P. D. Howe. Hurricane Risk Perceptions And Preparedness Among Florida Business
Owners. Masters, The Pennsylvania State University, 2009.
C. Huggel, A. Ka¨a¨b, W. Haeberli, P. Teysseire, and F. Paul. Remote sensing based
assessment of hazards from glacier lake outbursts: a case study in the Swiss Alps.
Canadian Geotechnical Journal, 39:316–330, 2002. ISSN 0008-3674. doi:
10.1139/t01-099.
T. Ide, J. Schilling, J. S. A. Link, J. Scheffran, G. Ngaruiya, and T. Weinzierl. On
exposure, vulnerability and violence: Spatial distribution of risk factors for climate
change and violent con flict across Kenya and Uganda. Political Geography, 43:
68–81, 2014. ISSN 0962-6298. doi: 10.1016/j.polgeo.2014.10.007.
IPCC. IPCC Emissions Scenarios. Technical report, Intergovernmental Panel on
Climate Change, 2000.
S. K. Jain, R. Keshri, A. Goswami, A. Sarkar, and A. Chaudhry. Identification of
droughtvulnerable areas using NOAA AVHRR data. International Journal of
Remote Sensing, 30(10):2653–2668, 2009. ISSN 0143-1161. doi:
10.1080/01431160802555788.
L. Ji and A. J. Peters. Assessing vegetation response to drought in the northern Great
Plains using vegetation and drought indices. Remote Sensing of Environment, 87(1):
85–98, sep 2003. ISSN 00344257. doi: 10.1016/S0034-4257(03)00174-3.
173
D. Johnson, Douglas Hamilton Anderson. The Ecology of Survival: Case Studies from
Northeast African History. IB Tauris, 1988.
B. Jones. Uganda: Why have aid agencies been in Karamoja for so long?, jan 2011.
URL https://tinyurl.com/jff3am3.
R. O. R. Juma. Turkana livelihood strategies and adaptation to drought in Kenya.
Thesis, Victoria University of Wellington, 2009. URL
http://researcharchive.vuw.ac.nz/handle/10063/1063.
S. Kagan, L. Pedersen, S. Ollech, and D. Knaute. The Karamoja Syndrome :
Transdisciplinary Systems Research Informing Policy and Advocacy, 2009. URL
http://www.cultura21.net/karamoja/docs/Karamoja syndrome.pdf.
A. Karnieli, N. Agam, R. T. Pinker, M. Anderson, M. L. Imhoff, G. G. Gutman,
N. Panov, and A. Goldberg. Use of NDVI and Land Surface Temperature for
Drought Assessment: Merits and Limitations. Journal of Climate, 23(3):618–633,
feb 2010. ISSN 0894-8755. doi: 10.1175/2009JCLI2900.1.
R. Kasasira. PM’s Office Suspends Tractor Project in Karamoja - Daily Monitor, apr
2013. URL https://tinyurl.com/zyse8wz.
B. T. Kassie, H. Hengsdijk, R. Ro¨tter, H. Kahiluoto, S. Asseng, and M. Van Ittersum.
Adapting to climate variability and change: Experiences from cereal-based farming
in the central rift and kobo valleys, Ethiopia. Environmental Management, 52(5):
1115–1131, 2013. ISSN 0364152X. doi: 10.1007/s00267-013-0145-2.
B. Kington. Historical Ethnography and the Collapse of Karamojong Culture:
Premature Reports of Trends. Technical Report June, African Studies Seminar
13.6.0, 2002.
A. Klisch and C. Atzberger. Operational drought monitoring in Kenya using MODIS
NDVI time series. Remote Sensing, 8(4):1–22, 2016. ISSN 20724292. doi:
10.3390/rs8040267.
F. N. Kogan. Droughts of the late 1980s in the United States as derived from NOAA
polar-orbiting satellite data. Bulletin of the American Meteorological Society, 76(5):
655–668, 1995. ISSN 0003-0007. doi: 10.1175/1520-0477.
F. N. Kogan. Global drought watch from space. Bulletin of the American
Meteorological Society, 78(4):621–636, 1997.
S. Kra¨tli. Karamoja with ’the rest of Uganda’. Nomadic Peoples, 14(2):3–23, 2010.
ISSN 08227942. doi: 10.3167/np.2010.140202.
S. Levine. What to do about Karamoja? Why Pastoralism is Not the Problem But the
Solution. Technical Report September, UN Food and Agricultural Organization,
2010.
174
P. D. Little, J. McPeak, C. Barrett, and P. Kristjanson. Challenging Orthodoxies:
Understanding Pastoral Poverty in East Africa. Development and Change, 39:
585–609, 2008.
W. Liu and F. N. Kogan. Monitoring regional drought using the vegetation condition
index. International Journal of Remote Sensing, 17(14), 1996.
X. Liu, Y. Wang, J. Peng, A. K. Braimoh, and H. Yin. Assessing vulnerability to
drought based on exposure, sensitivity and adaptive capacity: A case study in
middle Inner Mongolia of China. Chinese Geographical Science, 23(1):13–25, 2013.
ISSN 10020063. doi: 10.1007/s11769-012-0583-4.
A. Lotsch, M. A. Friedl, B. T. Anderson, and C. Tucker. Coupled
vegetation-precipitation variability observed from satellite and climate records.
Geophysical Research Letters, 30(14):8–11, 2003. ISSN 0094-8276. doi:
10.1029/2003GL017506.
J. Luzinda and J. Wilson. Soils of Southern Karamoja District, Uganda, 1959. URL
http://esdac.jrc.ec.europa.eu/content/soils-southern-karamoja-district-uganda.
T. J. Lybbert, Y. Kusunose, and N. Magnan. Drought Risk and Drought Response in
Morocco: Vulnerability, Risk Perceptions and Drought Coping among Rainfed
Cereal Farmers. In Agricultural & Applied Economics Association 2009 AAEA &
ACCI Joint Annual Meeting, 2009.
MAAIF. MAAIF Statistical Abstract. Technical report, Ministry Of Agriculture,
Animal Industry and Fisheries, 2010.
M. Mamdani. Karamoja: colonial roots of famine in NorthEast Uganda. Review of
African Political Economy, 9(25):66–73, 1982. ISSN 0305-6244. doi:
10.1080/03056248208703517.
M. N. Marshall. Sampling for qualitative research. Family practice, 13(6):522–5, dec
1996. ISSN 0263-2136. URL http://www.ncbi.nlm.nih.gov/pubmed/9023528.
K. Matthysen and S. Finardi. The Karamoja Cluster of Eastern Africa: Arms
Transfers and their Repercussions on Communal Security Perceptions, 2010.
J. J. McCarthy, O. F. Canziani, D. J. Leary, D. J. Dokken, and K. S. White. Climate
change 2001: impacts, adaptation, and vulnerability. In J. J. McCarthy, O. F.
Canziani, N. A. Leary, D. J. Dokken, and K. S. White, editors, Contribution of
Working Group II to the Third Assessment Report of the Intergovernmental Panel
on Climate Change., pages 1–1042. Cambridge University Press., 2001. ISBN 0 521
01500 6.
T. McKee, N. Doesken, and J. Kleist. The relationship of drought frequency and
duration to time scales. Proceedings of the 8th Conference on Applied Climatology:
American Meteorological Society, 17(20):6, 1993.
175
S. M. McNeeley, T. A. Beeton, and D. S. Ojima. Drought Risk and Adaptation in the
Interior United States: Understanding the Importance of Local Context for Resource
Management in Times of Drought*. Weather, Climate, and Society, 8(2):147–161,
apr 2016. ISSN 1948-8327. doi: 10.1175/WCAS-D-15-0042.1.
C. McSweeney, M. New, and G. Lizcano. UNDP Climate Change Country Profile.
Technical report, UNDP, 2015. URL http://country-profiles.geog.ox.ac.uk.
G. Metternicht, L. Hurni, and R. Gogu. Remote sensing of landslides: An analysis of
the potential contribution to geo-spatial systems for hazard assessment in
mountainous environments, 2005. ISSN 00344257.
E. Meze-hausken. Contrasting climate variability and meteorological drought with
perceived drought and climate change in northern Ethiopia. Climate Research, 27:
19–31, 2004.
M. Mirzeler and C. Young. Pastoral politics in the northeast periphery in Uganda:
AK-47 as change agent. The Journal of Modern African Studies, 38(3):407–429,
2000. ISSN 0022278X. doi: 10.1017/S0022278X00003402.
A. K. Mishra, A. V. M. Ines, N. N. Das, C. Prakash Khedun, V. P. Singh,
B. Sivakumar, and J. W. Hansen. Anatomy of a local-scale drought: Application of
assimilated remote sensing products, crop model, and statistical methods to an
agricultural drought study. Journal of Hydrology, 526:15–29, 2015. ISSN 00221694.
doi: 10.1016/j.jhydrol.2014.10.038.
M. Moritz. Commoditization and the Pursuit of Piety: The Transformation of an
African Pastoral System. PhD thesis, University of California, 2003.
M. Mortimore. Adapting to Drought: Farmers, Famines and Desertification in West
Africa. Cambridge University Press, 1989. ISBN 0002-7162.
M. Mortimore. Roots in the African Dust:Sustaining the sub-Saharan Drylands.
Cambridge University Press, 1998.
M. Mortimore. Dryland development: Success stories from West Africa. Environment,
47(1):8–21, 2005. ISSN 0013-9157. doi: 10.3200/envt.47.1.8-21.
J. F. Morton. Pastoralism, Drought and Planning. Lessons from Northern Kenya and
Elsewhere, 2001.
J. F. Morton. The impact of climate change on smallholder and subsistence agriculture.
Proceedings of the National Academy of Sciences of the United States of America,
104(50):19680–5, dec 2007. ISSN 1091-6490. doi: 10.1073/pnas.0701855104.
C. P. Mubaya, J. Njuki, E. P. Mutsvangwa, F. T. Mugabe, and D. Nanja. Climate
variability and change or multiple stressors? Farmer perceptions regarding threats to
livelihoods in Zimbabwe and Zambia. Journal of Environmental Management, 102:
9–17, 2012. ISSN 03014797. doi: 10.1016/j.jenvman.2012.02.005.
176
W. K. Mugerwa. Rangeland tenure and resource management: an overview of
pastoralism in uganda. Technical report, Makerere Institute of Social Research,
Kampala, Uganda, 1992.
H. Murad and A. K. M. S. Islam. Drought assessment using remote sensing and GIS in
north-west region of Bangladesh. International Conference on Water and Flood
Management, pages 861–877, 2011.
C. S. Murthy, B. Laxman, M. V. R. S. Sai, and P. G. Diwakar. Analysing agricultural
drought vulnerability at sub-district level through exposure, sensitivity and adaptive
capacity based composite index. International Archives of the Photogrammetry,
Remote Sensing and Spatial Information Sciences - ISPRS Archives, 40(8):65–70,
2014. ISSN 16821750. doi: 10.5194/isprsarchives-XL-8-65-2014.
R. J. Mwanga. Socio-political and Conflict Implications of Land Deals: Cases from
Kenya and Uganda. In M. G. Berhe, editor, A Delicate Balance: Land Use, Minority
Rights and Social Stability in the Horn of Africa, chapter 3, page 334. Institute for
Peace and Security Studies, Addis Ababa University, 2014. ISBN 9789994497546.
E. Mwebaza. A historical perspective of land problems in Uganda. The Effects of Brief
Mindfulness Intervention on Acute Pain Experience: An Examination of Individual
Difference, 1(17), 2015. ISSN 1098-6596. doi: 10.1017/CBO9781107415324.004.
B. Mwesigwa, D. N. Bushara, H. Tusiime, E. Twanza, S. Gule, J. Odong, A. Industry,
P. Athuhaire, A. Mpungu, E. Ssekimwanyi, J. Birantana, and D. Ocen. Karamoja
Food Security Assessment, 2014.
C. Nakalembe. Characterizing Agricultural Drought in the Karamoja sub-region of
Uganda with Meteorological and Satellite-Based Indices. ’in reveiw’, 2017.
C. Nakalembe, J. Dempewolf, and C. Justice. Agricultural land use change in
Karamoja Region , Uganda. Land Use Policy, 62:2–12, 2017. ISSN 0264-8377. doi:
10.1016/j.landusepol.2016.11.029. URL
http://dx.doi.org/10.1016/j.landusepol.2016.11.029.
National Drought Mitigation Center. The Standardized Precipitation Index, 2015.
URL
http://drought.unl.edu/portals/0/docs/spi-program-alternative-method.pdf.
National Environment Management Authority. The State of the Environment Report
for Uganda, 2004/05. Technical Report July, National Environmental Management
Authority, Kampala, Uganda, 2008.
NewVision. Govt to distribute relief food to Karamoja, sep 2015. URL
https://tinyurl.com/hkup8tm.
M. Niamir-Fuller. Managing Mobility in African Rangelands, 1999. URL
http://www.ilri.org/InfoServ/Webpub/fulldocs/PropertyRights/Chapter7.htm.
177
A. R. Nsibambi. National Integration in Uganda 1962-2013. Fountain Publishers, 2014.
OCHA. Focus on Karamoja. Technical Report 4, Office for the Coordination of
Humanterian Affairs (OCHA), 2009. URL
http://www.cultura21.net/karamoja/docs/FocusKja-SpecRep3.pdf.
Office of the Prime Minister. Karamoja Action Plan for Food Security (2009-2014).
Technical Report May, Office of the Prime Minister Uganda, Kampala, 2009. URL
http://opm.go.ug/manage/pdfs/disaster/ProgrammeDocumentKaramoja.pdf.
Office of the Prime Minister. The Karamoja Integrated Disarmament and
Development Programme. Technical Report January 2007, Office of the Prime
Minister, Kampala, Uganda, 2010. URL
http://www.ugandaclusters.ug/dwnlds/0204Karamoja/KIDDP.pdf.
Office of the Prime Minister. Karamoja Food Security Situation. Technical Report
September, Office of the Prime Minister, 2015.
J. S. Okonya, K. Syndikus, and J. Kroschel. Farmers’ Perception of and Coping
Strategies to Climate Change: Evidence From Six Agro-Ecological Zones of Uganda.
Journal of Agricultural Science, 5(8):252–263, 2013. ISSN 1916-9760. doi:
10.5539/jas.v5n8p252.
W. C. Palmer. Meteorological Drought. U.S. Weather Bureau Research Paper, 45:58,
1965.
A. J. Peters, E. A. Walter-Shea, L. Ji, A. Vin˜a, M. Hayes, and M. D. Svoboda. Drought
monitoring with NDVI-based standardized vegetation index. Photogrammetric
Engineering & Remote Sensing, 68(1):71–75, 2002. ISSN 0099-1112.
D. Pflugmacher, O. N. Krankina, W. B. Cohen, M. a. Friedl, D. Sulla-Menashe, R. E.
Kennedy, P. Nelson, T. V. Loboda, T. Kuemmerle, E. Dyukarev, V. Elsakov, and
V. I. Kharuk. Comparison and assessment of coarse resolution land cover maps for
Northern Eurasia. Remote Sensing of Environment, 115:3539–3553, 2011. ISSN
00344257. doi: 10.1016/j.rse.2011.08.016.
J. R. Porter, L. C. Xie, A. J., K. Cochrane, S. M. Howden, M. M. Iqbal, D. B. Lobell,
and M. I. Travasso. Food Security and Food Production Systems. In Climate
Change 2014: Impacts, Adaptation, and Vulnerability, chapter 7, pages 485–533.
Cambridge University Press., 2014.
J. Powell. Karamoja: A literature review. Technical report, SAFERWORLD,
Kampala, Uganda, 2010.
A. J. Prelog and L. M. Miller. Perceptions of Disaster Risk and Vulnerability in Rural
Texas. Journal of Rural Social Sciences, 28(3):1–31, 2013. ISSN 2151-4178. URL
http://www.ag.auburn.edu/auxiliary/srsa/pages/Articles/JRSS 2013
28/3/JRSS 2013 28 3 1-31.pdf.
178
M. D. Quam. Cattle Marketing and Pastoral Conservatism: Karamoja District,
Uganda, 1948-1970. African Studies Review, 21(1):49–71, 1978. ISSN 0002-0206.
doi: 10.2307/523763.
S. M. Quiring and S. Ganesh. Evaluating the utility of the Vegetation Condition Index
(VCI) for monitoring meteorological drought in Texas. Agricultural and Forest
Meteorology, 150(3):330–339, mar 2010. ISSN 01681923. doi:
10.1016/j.agrformet.2009.11.015.
M. R. Rahman, Z. H. Shi, and C. Chongfa. Soil erosion hazard evaluation - An
integrated use of remote sensing, GIS and statistical approaches with biophysical
parameters towards management strategies. Ecological Modelling, 220(1314):
1724–1734, 2009. doi: http://dx.doi.org/10.1016/j.ecolmodel.2009.04.004. URL
http://www.sciencedirect.com/science/article/pii/S0304380009002634.
T. Reardon and J. E. Taylor. Agroclimatic shock, income inequality, and poverty:
Evidence from Burkina Faso. World Development, 24(5):901–914, 1996. ISSN
0305750X. doi: 10.1016/0305-750X(96)00009-5.
L. W. Robinson. A complex-systems approach to pastoral commons. Human Ecology,
37(4):441–451, 2009. ISSN 03007839. doi: 10.1007/s10745-009-9253-2. URL
http://dx.doi.org/10.1007/s10745-009-9253-2.
R. J. Rowley, K. P. Price, and J. H. Kastens. Remote Sensing and the Rancher :
Linking Rancher Perception and Remote Sensing Remote Sensing and the Rancher :
Linking Rancher Perception and Remote Sensing. Management, 60(4):359–368,
2007. doi: 10.2111/1551-5028(2007)60.
M. A. Rugadya and H. Kamusiime. Tenure in mystery: the status of land under
wildlife, forestry and mining concessions in Karamoja Region, Uganda. Nomadic
Peoples, 17(1):33–65, 2013. ISSN 08227942. doi: 10.3167/np.2013.170103.
J. Sanyal and X. X. Lu. Remote sensing and GIS-based flood vulnerability assessment
of human settlements: a case study of Gangetic West Bengal, India. Hydrological
Processes, 19(18):3699–3716, nov 2005. ISSN 0885-6087. doi: 10.1002/hyp.5852.
D. Schro¨ter, M. J. Metzger, W. Cramer, and R. Leemans. Vulnerability assessment -
analysing the human-environment system in the face of global environmental change.
The ESS Bulletin, 2(2):11–17, 2004.
K. I. Simala and M. N. Amutabi. Small arms, cattle raiding and borderlands: The
Ilemi Triangle. In I. Abraham and W. van Schendel, editors, Illicit flows and
criminal things, chapter 7, pages 177–200. Indiana University Press, Bloomington,
2005. ISBN 025334669X.
A. Singh. Review Article Digital change detection techniques using remotely-sensed
data. International Journal of Remote Sensing, 10(6):37–41, jun 1989. ISSN
0143-1161. doi: 10.1080/01431168908903939.
179
M. F. W. Slegers. If only it would rain: Farmers’ perceptions of rainfall and drought in
semi-arid central Tanzania. Journal of Arid Environments, 72(11):2106–2123, nov
2008. ISSN 01401963. doi: 10.1016/j.jaridenv.2008.06.011.
V. U. Smakhtin and E. L. F. Schipper. Droughts: The impact of semantics and
perceptions. Water Policy, 10(2):131–143, 2008. ISSN 13667017. doi:
10.2166/wp.2008.036.
J. Stark. Climate Change and Conflict in Uganda : The Cattle Corridor and Karamoja.
Technical report, United States Agency for International Development, 2011.
E. Stites and D. Akabwai. Changing roles, shifting Risks: Livelihood impacts of
disarmament in Karamoja, Uganda. Technical report, Feinstein International
Center, 2009.
E. Stites, D. Akabwai, D. Mazurana, and P. Ateyo. Angering Akuju : Survival and
Suffering in Karamoja. Technical report, Feinstein International Center, 2007.
E. Stites, L. Fries, and D. Akabwai. Foraging and Fighting: Community Perspectives
on Natural Resources and Conflict in Southern Karamoja. Technical Report August,
Feinstein International Center and Save the Children, 2010.
M. B. Sundal. Difficult Decisions: Karimojong Healing in Conflict. PhD thesis,
University of Kansas, 2010.
K. Suwanprasert, S. Seto, and S. Kaewrueng. Integrated Drought Risk Indices from
Climate Based and Satellite-based Observation for Agricultural Drought Monitoring
in Thailand. Journal of Japan Society of Civil Engneers, 69(4):1 25–1 30, 2013.
J. Taylor, T. Stewart, and M. Downton. Perceptions of drought in the Ogallala Aquifer
region. Environment and Behavior, 20(2):150–175, 1988.
D. M. Tralli, R. G. Blom, V. Zlotnicki, A. Donnellan, and D. L. Evans. Satellite
remote sensing of earthquake, volcano, flood, landslide and coastal inundation
hazards. In ISPRS Journal of Photogrammetry and Remote Sensing, volume 59,
pages 185–198, 2005. ISBN 0924-2716. doi: 10.1016/j.isprsjprs.2005.02.002.
G. Trewartha. Tropical East Africa. In The Earth’s problem climates, chapter 9, page
334. University of Wisconsin Press, Madison, 1961.
C. J. Tucker. Red and photographic infrared linear combinations for monitoring
vegetation. Remote Sensing of Environment, 8(2):127–150, 1979. ISSN 00344257.
doi: 10.1016/0034-4257(79)90013-0.
C. G. Turvey and M. K. Mclaurin. Applicability of the Normalized Difference
Vegetation Index (NDVI) in Index-Based Crop Insurance Design. Weather, Climate,
and Society, 4:271, 2012. ISSN 1948-8327. doi: 10.1175/WCAS-D-11-00059.1.
180
UBOS. The Northern Uganda Baseline Survey (NUS), 2004. Technical report, Uganda
Bureau of Statistics, Kampala, Uganda, 2004. URL
http://www.ubos.org/unda/index.php/catalog/9.
UBOS. The Uganda National Panel Survey ( UNPS ) 2009-Documentation. Technical
report, Uganda Bureau of Statistics, Kampala, Uganda, 2009.
Uganda Bureau of Statistics. Population and Housing Census (PHC) 2002. Technical
report, Uganda Bureau of Statistics, 2002.
Uganda Bureau of Statistics. Uganda Census of Agriculture 2008/2009. Technical
report, Uganda Bureau of Statistics, Kampala, Uganda, 2010a.
Uganda Bureau of Statistics. Uganda census of agriculture 2008/2009, Volume IV:
Crop Area and Production Report, 2010b. URL www.ubos.org.
Uganda Bureau of Statistics. Uganda National Panel Survey 2011/2012 WAVE III
REPORT. Technical Report November, Uganda Bureau of Statistics, Kampala,
Uganda, 2013a.
Uganda Bureau of Statistics. Uganda National Panel Survey 2010/2011 Wave II
Report. Technical Report June, Uganda Bureau of Statistics, Kampala, Uganda,
2013b.
Uganda Bureau of Statistics. The Uganda National Panel Survey (UNPS) 2011/12:
Basic Information Document. Technical Report July, Uganda Bureau of Statistics,
Kampala, Uganda, 2014.
Uganda Bureau of Statistics. The National Population and Housing Census 2014 Main
Report. Technical report, Uganda Bureau of Statistics, Kampala, Uganda, 2016.
UNDP. Uganda Human Development Report 2007. Technical report, United Nations
Development Programme, 2007.
UNDP. Human Development Report 2015: Work for Human Development. Technical
report, United Nations Development Programme, New York, 2015.
United States Drought Monitor. U.S. Drought Monitor Classification Scheme, 2015.
URL http://droughtmonitor.unl.edu/aboutus/classificationscheme.aspx.
M. K. van Aalst, I. Burton, and T. Cannon. Community level adaptation to climate
change: The potential role of participatory community risk assessment. Global
Environmental Change, 18(1):165–179, 2008. ISSN 0959-3780. doi:
10.1016/j.gloenvcha.2007.06.002.
R. van Duinen, T. Filatova, P. Geurts, and A. van der Veen. Empirical Analysis of
Farmers’ Drought Risk Perception: Objective Factors, Personal Circumstances, and
Social Influence. Risk Analysis, 35(4):741–755, 2015. ISSN 15396924. doi:
10.1111/risa.12299.
181
C. J. van Westen, E. Castellanos, and S. L. Kuriakose. Spatial data for landslide
susceptibility, hazard, and vulnerability assessment: An overview. Engineering
Geology, 102(3-4):112–131, 2008. ISSN 0013-7952. doi: 10.1016/j.enggeo.2008.03.010.
E. Vermote, C. O. Justice, and F. M. Bre´on. Towards a generalized approach for
correction of the BRDF effect in MODIS directional reflectances. IEEE Transactions
on Geoscience and Remote Sensing, 47(3):898–908, 2009. ISSN 01962892. doi:
10.1109/TGRS.2008.2005977.
S. M. Vicente-Serrano. Evaluating the Impact of Drought Using Remote Sensing in a
Mediterranean, Semi-arid Region. Natural Hazards, 40(1):173–208, oct 2006. ISSN
0921-030X. doi: 10.1007/s11069-006-0009-7.
S. M. Vicente-Serrano, S. Beguer´ıa, and J. I. Lo´pez-Moreno. A Multiscalar Drought
Index Sensitive to Global Warming: The Standardized Precipitation
Evapotranspiration Index. Journal of Climate, 23(7):1696–1718, apr 2010. ISSN
0894-8755. doi: 10.1175/2009JCLI2909.1.
S. M. Vicente-Serrano, D. Cabello, M. Toma´s-Burguera, N. Mart´ın-Herna´ndez,
S. Beguer´ıa, C. Azorin-Molina, and A. E. Kenawy. Drought variability and land
degradation in semiarid regions: Assessment using remote sensing data and drought
indices (1982-2011). Remote Sensing, 7(4):4391–4423, 2015. ISSN 20724292. doi:
10.3390/rs70404391.
O. V. Wilhelmi and D. A. Wilhite. Assessing Vulnerability to Agricultural Drought: A
Nebraska Case Study. Natural Hazards, 25:37–58, 2002a. ISSN 0921030X. doi:
10.1023/A:1013388814894.
O. V. Wilhelmi and D. A. Wilhite. Assessing vulnerability to agricultural drought: A
Nebraska case study. Natural Hazards, 25(1):37–58, 2002b. ISSN 0921030X. doi:
10.1023/A:1013388814894.
D. Wilhite. Understanding the drought phenomenon: The role of definitions. Water
International, 10:111–120, 1985.
D. a. Wilhite, M. D. Svoboda, and M. J. Hayes. Understanding the complex impacts of
drought: A key to enhancing drought mitigation and preparedness. Water Resources
Management, 21(5):763–774, 2007. ISSN 09204741. doi: 10.1007/s11269-006-9076-5.
S. Wolfram and D. B. Lobell. Robust negative impacts of climate change on African
agriculture. Environmental Research Letters, 5(1):1–8, jan 2010. ISSN 1748-9326.
doi: 10.1088/1748-9326/5/1/014010. URL
http://stacks.iop.org/1748-9326/5/i=1/a=014010?key=crossref.6f26401c5082faeb2f666118e1e02b98
http://stacks.iop.org/1748-9326/5/i=1/a=014010.
World Food Program. Food Security and Nutrition Assessment. Technical Report
January, World Food Programme (WFP) & UNICEF, 2016.
182
World Food Programme. World Food Programme, 2017. URL
http://dataviz.vam.wfp.org/.
Y.-p. Xu, S.-j. Lin, Y. Huang, Q.-q. Zhang, and Q.-h. Ran. Drought analysis using
multi-scale standardized precipitation index in the Han River Basin, China. Journal
of Zhejiang University SCIENCE (Applied Physics & Engineering), 12(6):483–494,
jun 2011. ISSN 1673-565X. doi: 10.1631/jzus.A1000450.
S. E. Yang and B. F. Wu. Calculation of monthly precipitation anomaly percentage
using web-serviced remote sensing data. In Proceedings - 2nd IEEE International
Conference on Advanced Computer Control, ICACC 2010, volume 5, pages 621–625,
2010. ISBN 9781424458462. doi: 10.1109/ICACC.2010.5486796.
A. Zhang and G. Jia. Monitoring meteorological drought in semiarid regions using
multi-sensor microwave remote sensing data. Remote Sensing of Environment, 134:
12–23, jul 2013a. ISSN 00344257. doi: 10.1016/j.rse.2013.02.023.
A. Zhang and G. Jia. Monitoring meteorological drought in semiarid regions using
multi-sensor microwave remote sensing data. Remote Sensing of Environment, 134:
12–23, 2013b. ISSN 0034-4257. doi: 10.1016/j.rse.2013.02.023. URL
http://dx.doi.org/10.1016/j.rse.2013.02.023.
183