Content uploaded by Gina Ziervogel
Author content
All content in this area was uploaded by Gina Ziervogel on Dec 13, 2013
Content may be subject to copyright.
CLIMATE RESEARCH
Clim Res
Vol. 40: 133–146, 2009
doi: 10.3354/cr00804
Published December 10
1. INTRODUCTION
Climate change is expected to place considerable
additional stress on the biophysical, economic, political
and social systems that determine livelihood security
in Africa (Leary et al. 2008). Accordingly there is a
growing need for ‘anticipatory adaptation’ or proactive
rather than reactive management of climate change
risk. Successful anticipatory adaptation requires the
best available information concerning the nature of
future climate risks; therefore, it is vital that climate
change scenarios are used more effectively in adapta-
tion decision making.
Understanding how organizations are accessing and
using climate scenarios is a priority (McNabb & Sepic
1995, Hassol 2008). This is necessary for coordinating
climate science applications, as well as facilitating bet-
ter coordination of donor-funded activities (Nyong
2005, Patt et al. 2007, Wilby 2007). The present study
provides an assessment of the extent to which informa-
tion from climate change models is being integrated
into development practice and decision making in agri-
© Inter-Research 2009 · www.int-res.com*Email: gina@csag.uct.ac.za
Climate change scenarios and the development of
adaptation strategies in Africa: challenges and
opportunities
Gina Ziervogel
1, 2,
*
, Fernanda Zermoglio
1
1
Stockholm Environment Institute, Oxford Office, 266 Banbury Road, Suite 193, Oxford OX2 7DL, UK
2
Present address: Department of Environment and Geographical Science, University of Cape Town, Private Bag X3,
Rondebosch 7700, South Africa
ABSTRACT: Climate change is expected to intensify existing problems and create new combinations
of risks, particularly in Africa, where there is widespread poverty and dependence on the natural
environment. Accordingly, there is a growing need for proactive adaptation to climate change risks.
In order to achieve this, the requisite competence needs to be developed on the use and interpreta-
tion of climate information to support informed decisions. The present paper assesses the extent to
which climate change scenarios are currently used in developing adaptation strategies within the
agricultural development sector, with a focus on Africa. The data, based on interviews with practi-
tioners and donors working in the climate change field in Africa, suggest that although annual
climate information (such as seasonal climate forecasts) is used to a certain extent to inform and sup-
port some decisions, climate change scenarios are rarely used at present in agricultural development.
However, respondents suggest a number of ways to improve the application of climate change
science in these endeavors; these include strengthening technical skills for downscaling climate
models, as well as using scenario outputs to develop and prioritize robust locally relevant adaptation
strategies to provide examples of ‘good’ adaptation practice. Improved understanding, packaging,
and communication of climate scenarios are required between scientists, practitioners, policymakers
and civil society, both within areas in the global south as well as between the global south and north.
In addition, we argue that a paradigmatic shift is required from supply-driven activities to a user-
focused approach that addresses decision makers’ needs for climate change data. Such a shift would
focus on generating the information required to provide actionable suggestions to formulate viable
adaptation policies and reduce the negative consequences of climate change, particularly for Africa’s
most vulnerable groups.
KEY WORDS: Adaptation · Climate change science · Climate change data · Africa · Agriculture
Resale or republication not permitted without written consent of the publisher
Contribution to CR Special 20 ‘Integrating analysis of regional climate change and response options’
O
PEN
PEN
A
CCESS
CCESS
Clim Res 40: 133–146, 2009
cultural activities in Africa. Focusing on what is work-
ing and what is not working will help to steer future
research aimed at supporting adaptation to climate
change.
The paper begins by outlining the current availabil-
ity of climate change scenario information for the sup-
port of adaptation activities in the agricultural sector.
The methodological approach of the assessment is
then introduced, after which the findings present the
characteristics and different approaches of organiza-
tions using climate change information within the field
of agricultural development. The insights from the
interview material provide the context for the discus-
sion that highlights the barriers and opportunities that
need to be addressed so that climate scenarios can be
more effectively integrated into climate change adap-
tation in Africa.
2. LINKING CLIMATE SCIENCE TO AGRICULTURAL
ADAPTATION
Climate change refers to shifts in the mean state
of the climate or in its variability, persisting for an ex-
tended period (decades or longer). Adaptation strate-
gies and actions should aim to secure well-being in the
face of climate variability, climate change and a wide
variety of difficult to predict biophysical and social
contingencies. Adaptation is most relevant when it in-
fluences decisions that exist in light of climate change
and that have consequences on decadal time scales
(Stainforth et al. 2007). Annual climate information,
such as seasonal climate forecasts, has become widely
available, and much research has been done on the op-
portunities and barriers to using this information in
Africa (Ziervogel & Downing 2004, Vogel & O’Brien
2006, Washington et al. 2006, Patt et al. 2007). Yet, in
the complex nexus of what actions to take and at what
cost, the appropriate use of climate information in risk
assessments is still lacking at time scales longer than
decades, especially in the most vulnerable areas of the
world. Climate change scenarios provide information
relevant to planning beyond the next few decades, yet
are not widely used. The present paper focuses on why
this is so and what might be done to improve uptake.
Over 60% of Africans remain directly dependent on
agriculture and natural resources for their well-being
(FAO 2003). Agriculture is highly dependent on cli-
mate variability (Salinger et al. 2005), which is why the
threat of climate change is particularly urgent in Africa
(Boko et al. 2007). Despite the reliance of large num-
bers of the population on agriculture, agricultural
development has, historically, not been a priority of
governments, ≤1% of average national budgets going
to agriculture (FAO 2003). However, many donors and
non-governmental organizations (NGOs) have sup-
ported agriculture across the continent because of this
reliance and the potential to improve yields. The
Alliance for a Green Revolution in Africa (AGRA) is an
example of such an organization that is supporting
agricultural development from funding projects on
seeds and soils to markets and policies (www.agra-
alliance.org).
In spite of the documented exposure of agricultural
activities to projected changes in climate, few agricul-
tural organizations have made use of climate model
output. A sound agricultural risk-assessment process
should ensure that climate change is appropriately
taken into account in planning and decision-making
processes. On the basis of available data and informa-
tion, it is possible to analyze the conditions and trends
in climate parameters, from the most basic data (e.g.
maximum and minimum temperature and rainfall) to
more elaborate indicators (duration of the growing
season) to complex indices (satisfaction index of water
requirements for the growing season), in order to allow
the identification of important thresholds and trigger
points on short and medium time scales. This informa-
tion can be used to assess potential impacts and iden-
tify anticipatory adaptation measures, allowing the
responses available to be prioritized and compared
equitably with other risks, resource availability and
other decision-relevant issues. For example, only after
a thorough description of any hazard (e.g. droughts) is
obtained, can users evaluate the potential future
change in these conditions across results of global cir-
culation models. Historical patterns, biophysical con-
straints, socio-economic dynamics, adaptive capacity
and resilience; these are all components that must be
considered in determining the nature of adaptation
efforts required with respect to climate variability and
change. Describing conditions that can be tested in
projected results of global circulation models for future
changes first requires that these conditions are clearly
defined.
The direct application of GCM (global circulation
model) output has been relatively limited in impact
studies, due to their coarse spatial resolution (which
in many cases fails to adequately represent critical
regional variations), their temporal resolution (the ear-
liest future time period available for most GCMs —
2045 — is too far into the future for many users), the
computational and technical requirements of using
these data, and, until recently, difficulties with data
access (Hulme et al. 2001). This has led to a growing
recognition of the need for climate information at finer
scales (Robock et al. 1993, Wilby & Wigley 1997, Islam
et al. 2008, Pal et al. 2007, Stainforth et al. 2007).
Understanding the relationship between large-scale
climate signals and local or regional climate impacts is
134
Ziervogel & Zermoglio: Applying climate change scenarios in Africa
central to the volume of work aimed at downscaling
climate model information for local and regional deci-
sion makers. Downscaled output is often a key piece of
information for development practitioners, providing
an avenue through which they can evaluate the impact
of climate change on a specific site or region. Down-
scaling takes model output from GCMs and interprets
them (via statistical techniques such as simple regres-
sion or more complex methods like those available in
neural networks), in relation to local climate dynamics
(Hewitson & Crane 1996, Wilby & Wigley 1997, Wilby et
al. 1998, Giorgi et al. 2001, Tadross et al. 2005). Their
relationship to each other and modeling in general are
depicted in Fig. 1.
Two approaches dominate the downscaling efforts in
Africa, each grounded on a specific set of assumptions
and methodologies: dynamical downscaling (also known
as regional climate models; RCMs) and empirical down-
scaling. Dynamical downscaling is a computationally
and technically expensive method, a characteristic that
has limited the number of organizations employing this
approach (Rummukainen et al. 2004, Kay et al. 2006).
One example, PRECIS, a product of the UK Met Office,
until only recently relied exclusively on the GCMs
from Hadley Centre, with several other models now in
use, and is the most widely used downscaled model in
Africa.
Empirical downscaling, which is less computation-
ally intensive, makes use of quantitative relationships
between the state of the larger scale climatic environ-
ment and local variations sourced from historical data.
By coupling specific local observed climate data with
GCM output, it provides a valuable solution to over-
coming the mismatch in scale between climate model
projections and the unit under investigation, which is
often a much smaller area than that contained within
a grid cell (such as a drainage basin or a coastline).
Empirical downscaling can be applied to a grid or to a
particular meteorological station. The later sub-set of
empirical downscaling is more common and referred
to as statistical empirical downscaling.
The Climate Systems Analysis Group (CSAG, www.
csag.uct.ac.za), based at the University of Cape Town,
South Africa, operates one of the few empirical down-
scaled models used for the whole of Africa, which sim-
ulates responses to global climate change at a growing
number of meteorological station locations across the
African continent. The approach, based on self-organiz-
ing maps, explained in detail elsewhere (Hewitson &
Crane 2006), relates the daily atmospheric state, deter-
mined by historical NCEP (National Centers for Envi-
ronmental Prediction) re-analysis data, to probability
density functions that reflect the stochastic range of
local responses. GCM climate change simulations,
which project the change in daily atmospheric states,
can then be used in the downscaling procedure to pro-
ject the change in local climate. The approach delivers
daily and monthly climate change information, includ-
ing multiple rainfall and temperature parameters, for
multiple models and 2 future time horizons (2046 to
2065 and 2081 to 2100).
Another empirical downscaling approach, statistical
downscaling methodology (SDSM) has gained some
applicability to Africa (Wilby et al. 2002). SDSM applies
a single GCM model and requires the user to acquire
historical local climate data in order to run the down-
scaling. Much of the use of SDSM in Africa has
been restricted to students who have applied this ap-
proach in their masters and doctoral dissertations. For
example, a research project used SDSM to downscale
HadCM3 to estimate the impact of climate change on
Ethiopia’s Lake Ziway’s hydrology to 2099 (Zeray et al.
2006). Changes in climate variables
were applied to a hydrological model
so as to simulate future flows.
Existing adaptation studies and
programs have had limited aware-
ness of the availability of down-
scaled model output. This is gradu-
ally changing due to the progress
noted above, with a growing number
of recent studies benefiting from
the downscaled information pro-
vided by this and other downscaling
approaches (Wilby & Wigley 1997,
Prudhomme et al. 2003).
Nevertheless, organizations still
tend to reference climate change as
a backdrop to their work as opposed
to integrating probabilistic future
scenarios into their current planning
135
Observed
climate data
Impact models
Statistical
downscaling
Dynamic
downscaling
Grid
GCM
Station
Validation
Validation
Fig. 1. Overview of different types of climate models. Grid empirical downscaling
refers to the spatial interpolation of gridded GCM (global circulation model) and
RCM (regional climate model) output
Clim Res 40: 133–146, 2009
and research approaches. Distinguishing the docu-
mented cases of climate change adaptation in Africa
from general development and agricultural practice is
therefore a challenging task. The impacts of the East
African droughts, for example, have been countered in
some instances by digging and maintaining sand dams
at the bottoms of rivers. The dams allow for continued
cattle watering during dry periods and have reduced
cattle deaths and conflict. It is not possible, however,
to establish climate change as the trigger for the con-
struction of sand dams (or other adaptation measures),
and the people constructing sand dams do not draw on
climate change scenarios to target locations.
This apparent disconnection between adaptation ef-
forts and climate model output exists in spite of efforts
on behalf of climate scientists (including African Cen-
tre of Meteorological Applications for Development
[ACMAD], Centre Régional de Formation et d’Applica-
tion en Agrométéorologie et Hydrologie Opérationelle
[AGRHYMET] and CSAG at the University of Cape
Town, in Africa, and Walker Institute and Tyndall
Centre, in the UK) to make their work more relevant to
agriculture and the equally concerted attempts by the
impact community to clarify their climate information
needs. This can partly be attributed to a lack of under-
standing among agriculturalists of how climate model
output can or should be applied. Many farming groups
need to acknowledge how difficult it is to understand,
interpret and use climate information. It is equally
true, however, that much of the climate modeling work
remains focused on gaining greater understanding of
atmospheric dynamics and fails to appreciate the types
of issues confronted by farmers or the manner in which
model output needs to be packaged and communicated
so as to make it accessible to agricultural decision
makers (see Appendix 1).
3. METHODOLOGY
The study commenced with a literature and web
review to identify the key issues and organizations
working on climate change modeling and agricultural
development
1
. The organizational characteristics were
summarized in a database using the following cate-
gories: the geographic scale of operation, the organiza-
tion’s focus (whether crop development, agricultural
productivity, or a focus on food and livelihoods, for
example); whether or not they use data on current cli-
mate variability; whether or not they actively use cli-
mate change scenarios; and whether or not the institu-
tion has developed climate adaptation strategies.
The database facilitated clustering of organizations
according to their activities, geographic focus, scale
and use of climate data, which enabled identification
of the alignment of activities across stakeholders, as
well as potential duplications, a task which has been
identified as a priority according to previous studies
(Nyong 2005, Wilby 2007). Furthermore, the resulting
analysis allowed the research team to identify candi-
dates for the second phase of the study, which involved
in-depth interviews. Forty organizations were selected
2
across the disciplines of climate science, climate im-
pacts, vulnerability and adaptation, agricultural devel-
opment and climate variability. Semi-structured inter-
views were conducted with representatives of these
organizations in person and on the telephone between
2007 and 2008.
The questions aimed to evaluate the institution’s
current and planned climate change strategies,
strengths and adaptation pathways. A second objec-
tive was to understand how organizations currently
access and use climate model output, identifying
opportunities for improved delivery of information for
adaptation in Africa.
From Fig. 2b it is evident that the study focused on
specific donors addressing climate change adaptation.
Interestingly most donors are internationally based, as
are the majority of climate scientists. The national cli-
mate science activity that is taking place is linked to
organizations including ACMAD (based in Niamey,
Niger) and the ICPAC (IGAD [Intergovernmental
Authority on Development] Climate Prediction and
Application Centre based in Nairobi, Kenya). The
figure shows a lack of adaptation-specific practition-
ers. Although there are practitioners in the agricul-
tural sector who are exploring climate change impacts
and there are climate scientists focusing on adapta-
tion, there is a noteworthy lack of practitioners with
knowledge of both climate change and agricultural
adaptation.
136
1
Including individuals and organizations focused on agricul-
ture, a limited number of organizations supporting climate
adaptation, development-oriented donors and agencies, re-
searchers and climate scientists
2
The researchers attempted to interview most of the organiza-
tions in the database that appeared to have an interest or in-
volvement in both agriculture and climate issues, subject to
the availability of interviewees. It is recognized that some
organizations have presumably been excluded and that
involvement in climate change activities is likely to have
grown by the time the present paper is published, so the
selection is a sample. See Appendix 2 for the list of organisa-
tions interviewed
Ziervogel & Zermoglio: Applying climate change scenarios in Africa
4. FINDINGS: THE ROLE OF STAKEHOLDERS IN
CLIMATE CHANGE ADAPTATION IN AFRICA
4.1. The role of donors in supporting adaptation
It is clear that adaptation is becoming more widely
supported among donors. Government agencies are
scaling up their adaptation commitment, and the number
of people assigned to work on adaptation is growing. For
example, when interviewed the GTZ (German Agency
for Technical Cooperation) had 10 people working on cli-
mate change, with 2 dedicated to adaptation issues; the
USAID (United States Agency for International Develop-
ment) had 5 people working on climate change issues,
with 2 of those focused on adaptation; and the SIDA
(Swedish International Development Agency) had just
started to intensify their adaptation focus. The World
Bank had 15 people in its climate change group, 9 of
whom worked on adaptation globally. The United
Nations Food and Agricultural Organization (FAO), sim-
ilarly, had launched a well-attended Inter-Departmental
Working Group on Climate Change, with a specific
adaptation focus. The number of people and donors
working in this field is increasing steadily.
Many donors fund climate change adaptation work in
some way. At the national level, the UK Department for
International Development (DFID) fund a number of pro-
grams focusing on rural economies in Africa. The IDRC
(International Development Research Corporation of
Canada) and DFID have been supporting a £24 million,
5 yr program called Climate Change Adaptation in
Africa (CCAA). The DGIS (Netherlands Ministry of
Foreign Affairs/Directorate General for International Co-
operation) has assisted developing countries through the
Netherlands Climate Change Studies Assistance Pro-
gram (NCCSAP), which began in 1996; a second phase
(Netherlands Climate Assistance Program; NCAP)
started in 2003, with the aim of influencing policy. It was
active in 14 countries, including 5 countries in Africa.
At a multi-national level, the Global Environment
Facility (GEF) provides grants to developing countries
for projects that benefit the global environment and pro-
137
8
4
0
8
4
0
18
16
14
12
10
8
6
4
2
0
12
8
4
0
Climate
science
Climate
adaptation
Agricultural
development
Climate
science
Climate
adaptation
Agricultural
development
Climate
science
Climate
adaptation
Agricultural
development
Climate
science
Climate
adaptation
Agricultural
development
Donors
dc
Number of respondentsNumber of respondents
Research institutes
Number of respondents Number of respondents
Practitioners
Practitioners
Researchers
Donors and
development
agencies
Africa
Regional
International
ba
Primary focus Geographic focus
Fig. 2. Summary of stakeholder analyses. (a) Organizations were clustered according to their primary focus—climate science:
stakeholders or organizations producing or using climate science in their respective areas of work; climate adaptation: stake-
holders or organizations whose work incorporates or focuses on climate change adaptation; agricultural development: stake-
holders or organizations whose work focuses largely on agricultural development (organizations interviewed are given in
Appendix 2). (b,c,d) Summary of the mandates of respondents by cluster and the geographic focus of their work. Geographic
focus — Africa: national activities based within African countries; regional: activities across a region of Africa; international:
activities based outside of Africa
Clim Res 40: 133–146, 2009
mote sustainable livelihoods in local communities. The
GEF is an independent financial organization made up
of a collaboration between the UNEP (United Nations
Environment Program), the World Bank and the UNDP
(United Nations Development Program). Assessments of
Impacts and Adaptations to Climate Change (AIACC)
was one of the GEF’s key initial adaptation programs.
Donors are also recognizing the need to develop
material to guide adaptation to climate change pro-
jects. Both the GTZ and USAID have developed guid-
ance material to support the process of developing
adaptation projects and implementing risk manage-
ment strategies. Donors recognize the need to improve
the capacity to communicate climate information and
apply available climate information to responses and
are starting to address this need.
Although START (Global Change SysTem for Analy-
sis, Research and Training) is not a donor, they manage
a number of donor programs that relate to climate
change. The START representative highlighted a con-
cern that many development donors are supporting
flood, famine and conflict disaster responses and do
not see it as their responsibility to support long-term
research on climate science and adaptation, even
though this research could assist them in their prepara-
tion for disasters. At the same time, many science
agencies do not fund adaptation on the grounds that
they expect this area to be covered by development
donors. This leaves something of a gap in the available
funding for programs focused specifically on climate
change adaptation. As the START representative ob-
served in our interview with him:
‘It is important to adapt now but this should be sup-
ported by science, which requires science agencies
and donors to cooperate more’.
Whilst climate scientists seem to be generally ade-
quately funded, and traditional development agencies
are all aware of climate change, there are very few
programs specifically focused on linking climate sci-
ence with adaptation efforts
3
. Instead, most current
donor activities aim either to assess impacts and vul-
nerability to climate change, in order to determine pri-
orities, or to integrate a measure of ‘climate proofing’
4
in development projects.
Amongst the donors interviewed, agriculture was
often not their main focus. The GTZ acknowledged
that there has been a move away from focusing on
agriculture to focusing rather on issues connected with
nutrition, food quality and access to market. The SIDA
has explored how adaptation to climate change can be
integrated into poverty alleviation approaches within
different sectors. The USAID has supported some agri-
culturally focused activities and explored what addi-
tional challenges climate change may present. The
DFID maintains an agricultural focus in eastern and
southern Africa, but, at the request of West African
governments, has moved its focus in that region to
urban issues.
To some extent the gap in funding for adaptation-
specific activities is being filled by private sector fun-
ders. There has been a discernible increase in private
sector funding for development and climate change
activity. In the past 5 yr the activities of the Bill and
Melinda Gates Foundation, the Google Foundation
and the Rockefeller Foundation, among others, have
moved beyond conventional corporate social responsi-
bility and begun to engage systemically with global
issues.
4.2. The state of research around climate change
adaptation
Since the ratification of the United Nations Frame-
work Convention on Climate Change (UNFCCC), var-
ious international research programs have been initi-
ated in order to develop the capacity of developing
countries to cope with the effects of climate change.
Typically, these research programs are a collaborative
effort of many different organizations involving research
institutes, NGOs, as well as governmental organiza-
tions.
On the climate science side, the Intergovernmental
Panel on Climate Change (IPCC) is the preeminent
source of climate change information. The IPCC re-
views and collates research on a range of climate-
change-related issues and records climate data in its
data distribution centre (DDC). No African climate
research organizations are involved in producing
GCM models. Because of the lack of the necessary
tools, both human and instrumental, Africa depends, to
a very large extent, on organizations based in Europe
and North America for its operational climate forecast-
ing capacity. Three organizations —the NOAA (National
Oceanic and Atmospheric Administration, USA), UK
Met Office/Hadley Centre (UK) and the CNRS (Centre
National de la Recherche Scientifique, France) —rou-
tinely make seasonal climate forecasts for Africa based
on their respective GCM models. Among these 3
138
3
CGIAR (Consultative Group on International Agricultural
Research) has requested funding for a program and Michi-
gan State University’s CLIP (Climate-Land Interaction Pro-
ject) draws on local climate science
4
‘Climate proofing’ whilst in common use in the development
discourse, is not a term that the authors support on the
grounds that it is very difficult to predict all climate change
impacts and then secure projects and programs against cli-
mate change. Whilst most development organizations recog-
nize this, the term endures. ‘Climate preparedness’ might be
more appropriate
Ziervogel & Zermoglio: Applying climate change scenarios in Africa
institutions, the Hadley Centre, and specifically their
dynamically downscaled PRECIS model, is the most
widely consulted, especially in southern Africa.
Within Africa, a number of organizations engage in
climate work, although the need for regional modeling
and downscaling of GCM outputs is yet to be ade-
quately met from within Africa. The ICPAC, based in
Nairobi, Kenya, is a regional climate center that was
established in 1989 as the Drought Monitoring Centre
in Nairobi (DMCN) in an effort to minimize the nega-
tive impacts of extreme climate events like droughts
and floods along the Greater Horn of Africa. The major
goal of the ICPAC is to improve and enhance the pro-
duction and provision of sector-relevant climate in-
formation and applications in the region, focusing on
decadal (10 d), monthly and seasonal forecasts.
Whereas the focus of the center is primarily on the pro-
duction of information about climate variability, the
ICPAC has recently also started to work on generating
climate change projection data, mainly through the
use of the PRECIS dynamical downscaling model.
The ACMAD focuses on weather forecast informa-
tion and does not yet interpret climate predictions from
the model products. The CSAG is the only African
institution currently engaged in empirical downscaling
activities for climate change. In 2009 the CSAG made
available downscaled data from 10 GCM AR4 (IPCC
4th assessment report) models, including downscaled
daily data for the A2 scenario and for the 2046 to 2065
and 2081 to 2100 periods.
The National Center for Atmospheric Research
(NCAR) has used the MAGICC/Scengen software
package for future climate change prediction and, par-
ticularly, in drawing out the sources and magnitude
of uncertainty (see Wigley 2004). The downscaled cli-
mate model applies a regionalization algorithm. It
takes emissions scenarios for greenhouse gases, reac-
tive gases and sulfur dioxide as input and gives global
mean temperature, sea level rise and regional climate
as output. The software also quantifies the uncertain-
ties in these outputs. The regional results of the model
are based on results from 17 coupled atmosphere–
ocean general circulation models (AOGCMs), which
can be used individually or in any user-defined combi-
nation (Hulme et al. 2000).
One continental effort to support climate observa-
tions, climate risk management and climate policy
needs in Africa has been led by Global Climate
Observing System (GCOS). An initial report, funded
by the DFID and Defra (the UK Department for Envi-
ronment and Rural Affairs), produced prior to the UK
Presidency of the G8, reviewed Africa’s science gap
and the measures needed to reduce it. The report
showed that we know remarkably little about Africa’s
climate and that the climate-observing system for
Africa is the least developed of any of the continents
and is deteriorating. The level of technical expertise
available to support climate science in Africa, and,
hence, the level of activity, is very low (Wilby 2007:
p. 3). This process led to the emergence of ClimDev
(the Climate for Development in Africa program; de-
scribed further in the Section 4.3 below).
Research organizations focused on agriculture have
started to explore links to climate change. The CGIAR
(Consultative Group on International Agricultural Re-
search), with a secretariat in Washington, has an exclu-
sive agricultural focus (including forestry, fisheries,
aquaculture, water management), and the bulk of the
CGIAR’s climate change work is aimed at Africa, sup-
ported through the Challenge program. When inter-
viewed, they were developing climate change adapta-
tion programs. The FAO has a widespread presence in
Africa. They recently launched a climate change adap-
tation strategy, signaling internationally the FAO’s
intent to engage in climate change adaptation. The
FAO has access to a wealth of agri-met data, but does
not specifically apply climate change projections in its
agricultural support programs. The ICRISAT (Inter-
national Crops Research Institute for the Semi-Arid
Tropics), with regional hubs in Niger and Kenya, has
used regional IPCC data to explore the impact of a 3°C
warming on a variety of crops. Some of these IPCC
data were obtained from the download options avail-
able from the DDC.
There are a handful of organizations that are try-
ing to work across the science–society boundary. The
Stockholm Environment Institute (SEI), which spe-
cializes in sustainable development and environment
issues across scales, has supported and been
involved in numerous adaptation projects (Thomalla
et al. 2006, Klein et al. 2007, Ziervogel & Taylor
2008). The SEI Oxford center has recently been
working on a collaborative platform for climate adap-
tation, weADAPT, that includes climate scientists,
academics, practitioners and policymakers. This has
involved developing novel approaches to link risk of
future climate change, viewed within a multi-stressor
context, to robust adaptation actions, supported by a
range of stakeholders.
START International is another institution that has
worked across the sciences and social sciences by
supporting research activities in Africa, consistent
with critical environmental priorities on the continent,
which include food and water security and vulnera-
bility to climate change impacts (Leary et al. 2008).
START fosters regional networks of collaborating sci-
entists and organizations in developing countries to
conduct research on regional aspects of environmen-
tal change, assess impacts and vulnerabilities to such
changes, and to provide information to policy makers.
139
Clim Res 40: 133–146, 2009
START also provides a wide variety of training and
career development opportunities for young scien-
tists.
A more focused cross-disciplinary research approach
has been undertaken by Michigan State University
(MSU), which has a longstanding partnership with the
International Livestock Research Institute (ILRI), based
in Kenya. ILRI and MSU have conducted several joint
programs in East Africa looking at local interactions of
land use and climate (Smucker et al. 2007, Alagar-
swamy et al. 2008, Thornton et al. 2008). MSU’s flag-
ship adaptation project, CLIP (Climate-Land Interac-
tion Program), focuses on unraveling the complexities
of the climate system and the ways in which it is inter-
acting with the biophysical environment. Researchers
have developed a unique coupled modeling system
integrating human behavior and biophysical factors
that traces current and future land and climate change
in East Africa, allowing for geographically explicit
analyses of the impacts of land use and climate change
on natural and agricultural systems (Torbick et al.
2006, Olson et al. 2007).
Even though it is hard to generalize about the differ-
ent programs, a shift seems to have taken place in the
past 18 mo, from a primary focus on understanding the
nature of climate change towards the question who is
vulnerable and why. More recently, a further shift
has taken place towards increased attention to policy-
relevant applications and the question of how to adapt.
To some extent, these shifts are in line with the current
developments in the climate change debate, which has
gradually moved from a focus on awareness raising
about the reality of the problem towards a focus on
how to respond. Yet, the adaptation responses, whether
focused on policy or practice, are not yet integrating
climate science to the necessary extent.
4.3. Stakeholders involved at the grassroots level
The practitioners involved in actual adaptation ini-
tiatives appear to be fewer than those involved in
research, but a number of agricultural programs have
designed adaptation components.
United Nations Institute for Training and Research
(UNITAR) is managing the implementation of the
ACCCA (Advancing Capacity to Support Climate
Change Adaptation) program with the scientific sup-
port of a network of partners. Through research and
implementation, the ACCCA program seeks to support
effective adaptation decisions for the reduction of vul-
nerability to climate and environmental change. There
are projects in 19 countries, 14 of which are situated in
Africa. Building upon past experiences, the program
emphasizes the importance of developing strategies
for communicating climate risk information in a way
that is clear and understandable for policymakers.
ClimDev aims, among other goals, to improve the
availability, exchange and use of climate information
and services in support of economic growth and
achievement of the Millenium Development Goals
(MDGs), working at national, local and regional levels,
through improving climate observing networks and
services in Africa.
The FAO has conducted pilot studies in a number of
member countries, typically with local ministries of
agriculture. The pilot studies have various aims, but
broadly seek to combine better use of information with
improved technology and greater care for the natural
environment, so as to deliver enhanced and less risky
agricultural production. Whilst FAO pilot studies are
beginning to consider the implications of climate
change, they do not make systematic use of down-
scaled climate data in their planning.
SouthSouthNorth (SSN), which has operated as both
a local NGO (SSN Africa) and an international network
(SSN group), was initiated in 1999, but has recently
wrapped up its operations. Members worked on both
adaptation and mitigation projects. Other stakeholders
that fall under this category, but are not outlined in
detail include World Vision, ENDA-TM, ALMP (Arid
Lands Resource Management Project), AGRHYMET
and FEWSNET (Famine Early Warning System Net-
work).
In addition there are a number of private sector con-
sultancies active in the field of climate change adapta-
tion, and some NGOs now sell their services in this
field alongside consultancies on a wide range of mat-
ters including social learning, emissions management
and trading, bioenergy and capacity building. Exam-
ples include OneWorld Sustainable Investments, Envi-
ronmental Resource Management (ERM) and D1 oils.
5. DISCUSSION
Although there is growing awareness of and refer-
ences to climate change, much of these are based on
media messages, or highly aggregated model output
from the IPCC (DDC) and GCM models. The link
between climate change information and adaptation
practitioners on the ground remains largely non-
existent, and many adaptation practitioners in the agri-
cultural sector still rely on generalized assumptions
about how the climate will change or they derive very
general information about climate change and its
impacts from IPCC reports, as they are unaware of
where to access more detailed data and how to inter-
pret it. Those who do access model output tend to rely
on only one model creating the risk of drawing unsub-
140
Ziervogel & Zermoglio: Applying climate change scenarios in Africa
stantiated, or at best sub-optimal, inference. Decisions
are, more often than not, based not on that model’s
suitability to the research problem or adaptation chal-
lenge, but on its ease of access and use, with adoption
by simple word of mouth the extreme— but neverthe-
less plausible— case. Only rarely have researchers
provided a science-based rationale for choosing one
model over another. It is clear that there is a long way
to go before climate change scenarios are used effec-
tively in developing adaptation strategies.
Although some respondents view climate change data
as useful, other respondents suggested that there is a
lack of perceived relevance of climate model data for
agricultural stakeholders, which is why these data are
not used more widely. Similar challenges were experi-
enced when seasonal climate forecast information was
disseminated for agricultural use (Patt & Gwata 2002,
Vogel 2000, Archer 2003, Ziervogel 2004, Vogel &
O’Brien 2006). Agricultural decision makers require in-
formation on a range of matters in order to manage their
businesses and programs. Making the link between the
type of data that are reported by climate change models
(projections up to 2100) and the type of data that they
perceive as being important to their activities (market
demands, price, cost of inputs, labor availability, short-
term weather) is not always possible, or, in some in-
stances, is not considered a priority. Indeed, it posits a
great challenge that is difficult to overcome.
In order to monitor progress on how climate change
information is used, there needs to be a comprehensive
baseline that characterizes and contextualizes the
current adoption of climate change information in
Africa’s agricultural sector. Currently, contextualiza-
tion is hampered in 3 ways. First, there appears to be a
general tendency to isolate the climate change impacts
from the broader context in which developments are
taking place. There are several examples where cli-
mate change is assumed to be causing negative trends
without considering the possible importance of other
drivers. For instance, the ICRISAT respondent noted
an example from the Machakos district in Kenya,
where pastoralists blamed climate change and de-
creasing rainfall for decreasing crop yields, opting to
downplay the detrimental effects of overgrazing on
pasture resources. The meteorological records, how-
ever, indicated that rainfall had been increasing rather
than decreasing. This illustrates the importance of
putting climate change in a broader context, taking
into consideration other possible (and often more
directly probable) causes and explanations. Lack of
understanding of these other drivers of change might
unintentionally lead to misdirected projects and, in the
long run, to ‘maladaptation’.
Secondly, the vulnerability context is often not
understood. Respondents from both SEI and START,
who have provided technical support on climate vul-
nerability and adaptation projects in Africa, high-
lighted the tendency for organizations to adopt the
adaptation mandate without first clearly understand-
ing the climate change and vulnerability context. Many
organizations tend to take certain climate impacts and
vulnerabilities for granted without exploring which cli-
mate parameters and conditions are actually responsi-
ble for specific vulnerabilities to climate change and
how these parameters and conditions might change
under future scenarios (Ziervogel & Taylor 2008). This
might again lead to misdirected adaptation and devel-
opment measures.
Thirdly, climate change adaptation efforts often fail
to contextualize climate change risks within the set of
other climate information used in decision making,
including historical data, real-time data and traditional
knowledge, all of which are currently used and avail-
able to support decision-making processes. In fact,
there seems to be an apparent tension between people
working on future climate change and those focusing
on current climate variability. Some climate change
professionals argue that, although focusing on current
climate variability might equip agricultural decision-
makers over a short time horizon, they might then be
caught off guard by climate change, particularly
where the changes brought about by climate change
are significant and can be abrupt. In modeling par-
lance this problem is viewed as distinguishing the cli-
mate change ‘signal’ from the climate variability
‘noise’. In contrast, those who focus on climate variabil-
ity claim that unless farmers in Africa can be helped to
cope better with current climate variability, the chal-
lenge of adapting to future climate change will be
daunting for most and impossible for many. We argue
that these approaches are complimentary, rather than
mutually exclusive, and they must, eventually, be inte-
grated. A small-scale farmer, for instance, will be inter-
ested in seasonal climate forecasts that outline the
expected rainfall in the coming season in order to
make a decision on what crops to grow (Patt & Gwata
2002, Ziervogel et al. 2006). It should come as no sur-
prise that these farmers do not prioritize climate
change projections in their decision making. A crop
breeder, on the other hand, might benefit more from
an understanding of climate change patterns in the
next 20 to 30 yr, because of the time delay between the
development of new crops and the actual distribution
and use. In this case, an understanding of climate
change scenarios would be beneficial. Similarly a
donor agency looking to promote sustainable rural
development would be interested in climate-change-
induced shifts in agro-ecological zones over a 10 to
20 yr period, so that their current initiatives are not
undermined by future change.
141
Clim Res 40: 133–146, 2009
It is clear that, in addition to capturing baseline infor-
mation, there is a need to support increased develop-
ment and uptake of downscaled climate change projec-
tions and multi-model approaches in Africa. All climate
change modelers interviewed agreed that one of the
main barriers to producing climate change information
remains the lack of reliable meteorological data. This is
especially true for complex environments in which
higher concentrations of station data are needed to cap-
ture the complexity of the terrain. While many African
countries have established extensive monitoring net-
works in the course of the 20th century to support daily
weather forecasting, economic difficulties in the region,
not to mention civil wars, have led to the deterioration
of these networks in recent years. In fact, there are
fewer rainfall monitoring stations in many African
countries now than there were 20 or 30 yr ago (Wash-
ington et al. 2006). Ultimately, the lack of sufficient and
clean historical data renders the task of developing
sound and robust downscaling models difficult.
Another reason for the limited uptake of climate
change information is the lack of capacity, both in
terms of human resources and computational capacity,
to expand the available databases. Running dynamical
downscaling models requires considerable computa-
tional capacity. Currently CSAG and a handful of other
organizations (including ICPAC and ACMAD) have
some basic infrastructure and human resources in
place. More needs to be done to support theses insti-
tutes in their evolution into fully fledged climate mod-
eling centers.
Despite these barriers and shortcomings, the state of
climate science and modeling has reached a point
where it is able to adequately support local decision-
making processes. This point was supported by repre-
sentatives from START, SIDA, DFID and CSAG. The
critical focus, however, should be placed on the devel-
opment of sector-specific methods and examples of
how the climate model output could be utilized to sup-
port robust adaptation responses. The modeling com-
munity needs to focus on expanding the modeling
efforts within Africa, while working closely and inter-
actively with the users of model information in the
interpretation and understanding of climate model
output.
For farmers and other agricultural decision makers,
there are costs and risks involved when modifying
their age-old activities and practices in order to adapt
to what models indicate will happen. Some farmers
and program operators noted that it makes more sense
to react to observed (or historical) changes in weather
than to alter their activities based on a predicted cli-
mate risk. It is further true that many decision makers
are unable to contextualize the uncertainty that is
inherent in climate projections and, therefore, stick to
what they know. The reality is, however, that most
models concur on the near-term direction of change. If
used with the correct caution, these models can pro-
vide a sound, scientifically grounded basis for decision
making. The challenge comes in the timing of the
adaptation response, as it is unlikely that the models
will be able to provide enough information on when
the threshold will be crossed; thus, farmers will need to
decide on timing themselves.
Increasing uptake is going to require bridging the
gap between what scientists produce and what end-
users require. The DFID representative pointed out
that the majority of climate model data available have
been produced without due consideration to the spe-
cific (and admittedly diverse) needs of the people that
might need to use them and, therefore, is packaged in
a way that makes it difficult for practitioners to utilize.
A farmer, for example, may be less interested in mean
annual temperature, but would be very interested in
knowing how many years in an orchard life-time of say
25 yr sub-optimal rainfall or extreme temperatures
might be expected, and how this number of years can
be expected to change. This might be impossible to
infer from a reported temperature increase, especially
when the relationship between extremes and averages
is not established.
Bridging the gap between climatologists and end-
users will require augmentation of the people and orga-
nizations able to interpret and communicate this infor-
mation effectively. These translation skills (of so-called
boundary organizations) are necessary in order to en-
gage a wide range of stakeholders with specific needs,
as the limited number of climate scientists in Africa are
unable to develop their science at the same time as
meeting the growing need for model output interpreta-
tion and communication (Vogel et al. 2007).
At the same time, input from African scientists is
needed to drive the climate change agenda. Currently,
the growth in climate adaptation and agricultural de-
velopment activity in Africa has been mobilized from
outside of Africa, rather than from within. The GTZ
respondent suggested that projects have emerged in
response to foreign funding and so are ‘supply driven’.
He emphasized the need to be more focused on ‘de-
mand driven’ approaches that address locally identi-
fied needs. It does not have to be one or the other, but
rather flexibility is needed to enable demand-driven
approaches to emerge alongside supply-driven ap-
proaches and in doing so promote buy-in (i.e. accep-
tance) from a wider local community.
In conclusion, as adaptation rises in prominence on
the international agenda, so does the need to increase
and, in many cases, develop the requisite competency
for the use and interpretation of climate change sce-
narios to support informed decisions. A particular chal-
142
Ziervogel & Zermoglio: Applying climate change scenarios in Africa
lenge for those tasked with the design and implemen-
tation of adaptation projects is in leveraging the best
available data and synthesis tools to understand how
the expected climatic changes will exacerbate or induce
vulnerability of different activities under a changing
climate. To date, few tools or exercises provide the
much needed climate change information relevant to
decision makers. Most discouraging is the paucity of
work conducted at the nexus between climate scien-
tists and those concerned with making decisions. It is
against this background and with the ultimate goal of
improving the understanding, packaging, delivery and
communication of climate change scenarios that the
present study intended to set a process in motion that
will encourage the exchange of information among
providers of climate data and users (Ward 2008).
The present study offers valuable insight into the
challenges and opportunities for using climate change
scenarios in agricultural decision making in Africa.
Lessons from agricultural development stakeholders
are likely to be relevant to other sectors, although
further exploration is needed to establish the extent to
which stakeholders in other sectors are using climate-
change scenario information. Sustained communica-
tion and use of the data across organizations and disci-
plines will contribute to the development of linkages
between information users and data providers in tar-
geted ways. The authors expect that the analysis pre-
sented will contribute to the required paradigmatic
shifts: from supply-driven activity to a user-focused
understanding of the needs of decision makers accord-
ing to long-term climate data. Providing data relevant
to decision makers will support the development of
appropriate climate change adaptation practice and
policy that is particularly urgent for Africa’s most vul-
nerable groups.
Acknowledgements. The present study was based on the
report ‘Climate Change and Adaptation in African Agricul-
ture’ (2008), prepared for the Rockefeller Foundation by G.
Ziervogel, A. Cartwright, A. Tas, J. Adejuwon, F. Zermoglio,
M. Shale and B. Smith on behalf of the Stockholm Environ-
ment Institute. The present paper builds on foundations set at
the IPCC TGICA Meeting in Fiji in 2007, where significant
progress was made in discussions between climate scientists
and those involved in the vulnerability and adaptation field.
The authors thank 2 anonymous reviewers and Anthony Patt
for their valuable comments.
LITERATURE CITED
Alagarswamy G, Moore N, Lofgren B, Pijanowski B and
others (2008) Food production impacts of climate change
and land use change in East Africa. EOS Trans Amer
Geophys Union 89(53), Fall Meet Suppl, Abstract
GC53B-0730
Archer ERM (2003) Identifying underserved end-user groups
in the provision of climate information. Bull Am Meteorol
Soc 84:1525–1532
Benson C, Clay EJ (1998) The impact of drought on sub-
Saharan African economies: a preliminary examination.
Tech Pap 401, World Bank, Washington, DC
Boko M, Niang I, Nyong A, Vogel C and others (2007) Africa.
In: Parry ML, Canziani OF, Palutikof JP, van der Linden
PJ, Hanson CE (eds) Climate change 2007: impacts, adap-
tation and vulnerability. Contribution of Working Group II
to the 4th assessment report of the Intergovernmental
Panel on Climate Change. Cambridge University Press,
Cambridge, p 433–467
FAO (Food and Agricultural Organisation of the United
Nations) (2003) Responding to agricultural and food inse-
curity challenges mobilising Africa to implement Nepad
Programmes. Conference of ministers of agriculture of the
African union, Maputo
Gadain HM, Funk C (2003) Data and techniques to assist
RVF/forage model development. USGS, Nairobi
Giorgi F, Hewitson B, Christensen J, Hulme M and others
(2001) Regional climate information —evaluation and pro-
jections. In: Houghton JT, Ding Y, Griggs DJ, Noguer M,
van der Linden PJ, Dai X, Maskell K Johnson CA (eds)
Climate change 2001: the scientific basis. Contribution
of Working Group I to the 3rd assessment report of the
Intergovernmental Panel on Climate Change. Cambridge
University Press, Cambridge, p 583–638
Hassol SJ (2008) Improving how scientists communicate
about climate change. EOS 89(11):106–107
Hewitson BC, Crane RG (1996) Climate downscaling tech-
niques and application. Clim Res 7:85–95
Hewitson BC, Crane RG (2006) Consensus between GCM
climate change projections with empirical downscaling.
Int J Climatol 26:1315–1337
Hudson J, Vogel C (2003) The use of seasonal forecasts by
livestock farmers in South Africa. In: O’Brien K, Vogel C
(eds) Coping with climate variability: the use of seasonal
forecasts in southern Africa, Chap 1. Ashgate Publishing,
Aldershot, p 75–96
Hulme M, Brown O (1998) Portraying climate scenario uncer-
tainties in relation to tolerable regional climate change.
Clim Res 10:1–14
Hulme M, Wigley TML, Barrow EM, Raper SCB, Cantella A,
Smith S, Chipanshi AC (2000) Using a climate scenario
generator in vulnerablility and adaptation assessments:
MAGICC and SCENGEN v.2.4 workbook. Climatic
Research Unit, University of East Anglia, Norwich
Hulme M, Doherty R, Ngara T, New M, Lister D (2001)
African climate change: 1900–2100. Clim Res 17:145–168
IPCC (Intergovernmental Panel on Climate Change) (2007)
Summary for policymakers. In: Parry ML, Canziani OF,
Palutikof JP, van der Linden PJ, Hanson CE (eds) Climate
change 2007: impacts, adaptation and vulnerability. Con-
tribution of Working Group II to the 4th assessment report
of the Intergovernmental Panel on Climate Change. Cam-
bridge University Press, Cambridge, p 7–22
Islam MN, Rafiuddin M, Ahmed AU, Kolli RK (2008) Calibra-
tion of PRECIS in employing future scenarios in
Bangladesh. Int J Climatol 28:617–628
Johnston PA, Archer ERM, Vogel CH, Bezuidenhout CN, Ten-
nant WJ, Kuschke R (2004) Review of seasonal forecasting
in South Africa: producer to end-user. Clim Res 28:67–82
Kay AL, Reynard NS, Jones RG (2006) RCM rainfall for UK
flood frequency estimation. I. Method and validation.
J Hydrol (Amst) 318:151–162
Klein RJT, Eriksen SEH, Næss LO, Hammill A, Robledo C,
O’Brien KL, Tanner TM (2007) Portfolio screening to sup-
143
➤
➤
➤
➤
➤
➤
➤
➤
Clim Res 40: 133–146, 2009
port the mainstreaming of adaptation to climate change
into development assistance. Clim Change 84:23–44
Leary N, Adejuwon J, Barros V, Batimaa P and others (2008)
A stitch in time: general lessons from specific cases. In:
Leary N, Adejuwon J, Barros V, Burton I, Kulkarni I (eds)
Climate change and adaptation. Earthscan Publications,
London, p 1–28
McNabb DE, Sepic FT (1995) Culture, climate and total qual-
ity management: measuring readiness for change. Public
Productivity and Management Review 18(4):369–382
Nyong T (2005) Report on Africa-wide consultation on DFID/
IDRC proposed collaborative research and capacity devel-
opment program on climate adaptation research. Centre
for Environmental Resources and Hazards Research,
Department of Geography and Planning, University of
Jos, Nigeria. Available at: www.idrc.ca/uploads/user-S/
11416574281Nyong_Africa-Wide_consultation.doc
O’Brien KL, Vogel HC (eds) (2003) Coping with climate vari-
ability: the use of seasonal climate forecasts in southern
Africa. Ashgate Publishing, Aldershot
Olson J, Alagarswamy G, Andresen J, Campbell D and others
(2007) Integrating diverse methods to understand climate–
land interactions in East Africa. Geoforum 39:898–911,
doi:10.1016/j.geoforum.2007.03.011
Patt AG, Gwata C (2002) Effective seasonal climate forecast
applications: examining constraints for subsistence farm-
ers in Zimbabwe. Glob Environ Change 12:185–195
Patt AG, Ogallo L, Hellmuth M (2007) Learning from ten years
of climate outlook forums in Africa. Science 318:49–50
Prudhomme C, Jakob D, Svenson C (2003) Uncertainty and
climate change impact on the flood regime of small UK
catchments. J Hydrol (Amst) 277:1–23
Robock A, Turco R, Harwell M, Ackerman TP, Andressen R,
Chang HS, Sivakumar MVK (1993) Use of general circula-
tion model output in the creation of climate change sce-
narios for impact analysis. Clim Change 23:293–335
Rummukainen M, Bergstrom S, Persson G, Rodhem J, Tjern-
strom M (2004) The Swedish regional climate modelling
programme, SWECLIM: a review. Ambio 33:176–182
Salinger J, Sivakumar MVK, Motha RP (2005) Reducing vul-
nerability of agriculture and forestry to climate variability
and change: workshop summary and recommendations.
Clim Change70:341–362
Smucker T, Campbell D, Olson J, Wangui E (2007) Contempo-
rary challenges of participatory field research for land use
change analyses: examples from Kenya. Field Methods 19:
384–416
Stainforth DA, Downing TE, Washington R, Lopez A, New M
(2007) Issues in the interpretation of climate model
ensembles to inform decisions. Philos Trans R Soc 365:
2163–2177
Tadross MA, Jack C, Hewitson BC (2005) On RCM-based
projections of change in southern African summer climate.
Geophys Res Lett 32:L23713
Thomalla F, Downing TE, Spanger-Siegfried E, Han G, Rock-
ström J (2006) Reducing hazard vulnerability: towards a
common approach between disaster risk reduction and
climate adaptation. Disasters 30:39–48
Thornton PK, Jones PG, Alagarswamy G, Andresen J (2008)
Spatial variation of crop yield response to climate change
in East Africa. Global Environ Change 19:54–65, doi:
10.1016/j.gloenvcha.2008.08.005
Torbick N, Qi J, Lusch D, Olson J, Moore N, Ge J (2006)
Developing land use/land cover and parameterization for
climate and land modeling in East Africa. Int J Remote
Sens 27:4227–4244
Vogel C (2000) Usable science: an assessment of long-term
seasonal forecasts amongst farmers in rural areas of South
Africa. S Afr Geogr J 82:107–116
Vogel C, O’Brien K (2006) Who can eat information? Examining
the effectiveness of seasonal climate forecasts and regional
climate-risk management strategies. Clim Res 33:111–122
Vogel C, Moser SC, Kasperson RE, Dabelko GD (2007) Link-
ing vulnerability, adaptation, and resilience science to
practice: pathways, players, and partnerships. Glob Envi-
ron Change 17:349–364
Ward B (2008) Communicating on climate change, an essen-
tial resource for journalists, scientists and educators. Met-
calf Institute for Marine and Environmental Reporting,
Narragansett, RI
Washington R, Harrison M, Conway D, Black E and others
(2006) African climate change: taking the shorter route.
Bull Am Meteorol Soc 87:1355–1366
Wigley TML (2004) MAGICC/SCENGEN. Colorado National
Center for Atmospheric Research, Boulder, CO. Available
at: www.cgd.ucar.edu/cas/wigley/magicc/
Wilby R (2007) Decadal climate forecasting techniques for
adaptation and development planning: a briefing docu-
ment on available methods, constraints, risks and opportu-
nities. Department for International Development, UK
Wilby R, Wigley TML (1997) Downscaling general circulation
model output: a review of methods and limitations. Prog
Phys Geogr 21:530–548
Wilby RL, Wigley TML, Conway D, Jones PD, Hewitson BC,
Main J, Wilks DS (1998) Statistical downscaling of general
circulation model output: a comparison of methods. Water
Resour Res 34:2995–3008
Wilby R, Dawson CW, Barrow EM (2002) SDSM—a decision
support tool for the assessment of regional climate change
impacts. Environ Model Softw 17:145–157
Zeray L, Roehrig J, Chekol DA, Tropentag (2006) Prosperity
and poverty in a globalised world —challenges for agri-
cultural research: climate change impact on Lake Ziway
watershed’s water availability, Proc Tropentag Conf on Int
Agric Res Dev, October 11–13, 2006. University of Bonn
Ziervogel G, Downing TE (2004) Stakeholder networks:
improving seasonal forecasts. Clim Change 65:73–101
Ziervogel G (2004) Targeting seasonal climate forecasts for
integration into household level decisions: the case of
smallholder farmers in Lesotho. Geogr J 170:6–21
Ziervogel G, Taylor A (2008) Feeling stressed: integrating
climate adaptation with other priorities in South Africa.
Environment 50:32–41
Ziervogel G, Bharwani S, Downing TE (2006) Adapting to cli-
mate variability: pumpkins, people and policy. Nat Resour
Forum 30:294–305
144
➤
➤➤
➤
➤
➤
➤
➤
➤➤
➤
➤
➤
➤
➤
➤
➤
➤
Ziervogel & Zermoglio: Applying climate change scenarios in Africa
145
An important component of climate change science
involves the description, understanding and representation
of the inherent uncertainties in the modeling efforts (Stain-
forth et al. 2007). Whilst it is known that some models are
more ‘skilled’ at predicting specific parameters in certain
regions, without a comprehensive exploration of multiple
model outputs choosing a single model for a specific region
is not advisable (IPCC 2007). An analysis of a large set of
models (also referred to as an ‘ensemble’ of models), rather
than a single model, is currently the most robust approach
applied to addressing the uncertainty inherent in making a
decision that is influenced by the future evolution of the cli-
mate system, which is not entirely known or understood.
These ‘envelopes’—as the range of multiple model outputs
are called —of climate change help define the climatologi-
cal boundaries of potential climate change from a wide
range of scenarios and ought to include an exploration of
various scientific uncertainties. They are driven by the search
for climate spaces that are relevant to the needs of specific
localities and sectors (Hulme & Brown 1998, Stainforth et al.
2007).
Whilst envelopes do not provide the type of discrete
answers that some decision makers seek, they do offer deci-
sion makers with valuable perspectives on what might be
expected, which, in turn, can be used to attach the appropri-
ate level of confidence to models and associated decisions.
Wrongly assumed confidence levels in climate scenarios can
be as dangerous (sometimes more dangerous) than having
no projection at all. Fig. A1 provides an example of the
envelope approach by considering 7 different downscaled
precipitation anomaly scenarios for Bougouni, Mali. The
outer range (light blue and purple lines) of the combined
model projections provides the ‘envelope’ within which pre-
cipitation is expected to change in the early 21st century (to
2065). As such, it is the best estimate of expected future
changes.
It is in large part due to the issues outlined above that
adaptation activities in the agricultural sector in Africa have
focused more on perceived climate variability rather than
climate change. Seasonal forecasting and drought early
warning systems are increasingly used (especially among
livestock farmers) on the continent (Benson & Clay 1998,
Gadain & Funk 2003, Hudson & Vogel 2003, O’Brien &
Vogel 2003, Johnston et al. 2004, Ziervogel et al. 2006, Patt
et al. 2007). Whilst responding to historical change has
proven worthwhile, there is a danger that farmers, agricul-
tural policymakers, crop breeders and government officials
who structure their activities around short-term climate
variability will be caught unaware by the trends in climate
change and their longer term implications.
Appendix 1. Uncertainty and interpretation in climate science
Change in rainfall (mm)
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
Range of Models (‘Envelope’)
CCMA_CGCM3.1
CNRM_CM3
CSIRO_MK3.5
GFDL_CM2_0
GISS
IPSL_CM4
MIUB_ECHO_G
MPI_ECHAM5
MRI_CGCM2.3.2a
0
20
40
60
–80
–60
–40
–20
0
Month
Fig. A1. Average monthly projected changes in rainfall, Bougouni, Mali (2045–2065) by model. Different colors represent dif-
ferent model projections according to the key; 2 additional curves (triangle symbols) represent the envelope range of the out-
put. This output, from the Climate Change Explorer developed as part of the weAdapt platform (www.weADAPT.org), indi-
cates model agreement in the changing distribution of rainfall from May through September, which comprises the growing
season for this region in Mali, but less agreement for the start of the season in March. The graph clearly indicates that the
changes are not uniform across the season, with agreement among models concerning the expected increase in rainfall in the
middle of the year
Clim Res 40: 133–146, 2009146
Climate science
J. K. University, Kenya (ACCCA project —Advancing Capac-
ity to Support Climate Change Adaptation); Federal Univ.
Technology, Akure, Nigeria; Climate Systems Analysis
Group, University of Cape Town; Michigan State Univer-
sity; SDSM (Statistical Downscaling Methodology); Tyndall
Centre; Walker Institute for Climate Systems research;
GCOS (Global Climate Observing System); Zambia Meteo-
rological Services; ICPAC (IGAD Climate Predictions and
Applications Centre); ACMAD (African Centre of Meteoro-
logical Applications for Development); Nigeria Ministry of
Environment (special climate change unit).
Climate adaptation
GTZ (German Agency for Technical Cooperation); SIDA
(Swedish International Development Agency); USAID (United
States Agency for International Development); DGIS (The
international cooperation department of the Netherlands
Ministry of Foreign Affairs); DFID (UK Department for Inter-
national Development); ACCCA project Burkina Faso;
ACCCA project Malawi; Zambian Red Cross Society; Min-
istry of Environment and Sanitation, Mali; CLIP (Climate-
Land Interaction Project) and EACLIPSE (ILRI); CCAA (Cli-
mate Change Adaptation in Africa); LEAD Network; START
(Global Change System for Analysis, Research and Train-
ing); SEI (Stockholm Environment Institute); IIASA (Inter-
national Institute for Applied Systems Analysis); IIED
(International Institute for Environment and Development);
SSN (SouthSouthNorth) Africa, SSN Group.
Agricultural development
DFID; Buea University (ACCCA) project; Cameroon; Lake
Chad Research Institute; University of Pretoria, Economic
Department; Centre for Arid Zones Study; IAR&T (Institute
for Agricultural Research and Training); IITA (International
Institute for Tropical Agriculture); AGRA (Alliance for a
Green Revolution in Africa); AERC (African Economic
Research Consortium); CGIAR (Consultative Group on
International Agricultural Research); ICRAF (International
Centre for Research in Agro Forestry); ALMP (Office of the
President, Ministry of State for Special Programmes, Arid
Lands Resource Management Project); Jabenzi; World
Vision; FAO (UN Food and Agriculture Organization);
FEWSNET (Famine Early Warning System Network);
AGRHYMET (Centre Régional de Formation et d’Applica-
tion en Agrométéorologie et Hydrologie Opérationelle).
Appendix 2. Organizations interviewed
Submitted: July 23, 2008; Accepted: May 12, 2009 Proofs received from author(s): September 29, 2009
