Information Technology Reference
In-Depth Information
I would not believe you. There will be bad droughts again but only God knows what
will happen in the futureā.
Also, participants were asked about future scenarios to guide questions for later
versions of the model. For example, they were asked what they think will change in
their area in the next 5-10 years, will there be more or less rain and will more or fewer
people be farming when their children are herders, and so on. Based on the scenarios
identified by the participants, future versions of the model should include a field of
different agents which influence the motivations and options of the primary herder
agents, incorporating important power dynamics and wider political and socio-
economic processes. The factors affecting possible land-use changes, such as privati-
sation and sale of communal land, expansion of protected areas and spread of
irrigation should also be explored.
A limitation of using qualitative methods to inform ABM development is that nu-
merical thresholds and values often have to be applied, which can add rigidity beyond
the evidence presented in that data [4]. What became evident in talking to participants
was that decision-making processes are heuristic and qualitative. There are no set
thresholds or criteria for particular decisions, rather daily or seasonal assessments
based on the best available information and individual beliefs. For example, when
asked about decisions of when to increase their herds, most participants said they
generally prefer to buy stock just before the rains arrive when prices are low and fo-
rage not widely available, but this has to be weighed against the likelihood of a good
rainy season to support extra stock and rising market prices after the rains arrive. In
this sense, there is no set time or formula for decisions to go to market. For this rea-
son, actual set figures and numbers have not been used in the ABM rather ranges to
represent beliefs about rainfall for example, thus capturing the spaces within which
decisions are made relative to other factors rather than linking decisions to specific
variables like market prices.
5
Conclusions and Further Work
Qualitative methods have been used to capture narratives of livelihoods, climate
change and land-use change in a dryland social-ecological system in southern Kenya.
Building an ABM to explore these narratives offers a means of focussing on key
processes and their implications for the system while incorporating the agency of
actors to respond to change. Agent-based modelling also allows for the inclusion of
material and normative aspects in system analysis. Using qualitative data to inform
behavioural rules maximises this potential, especially when participants can be in-
volved in model-building and interpretation to capture a more
emic
representation of
the system which reflect the perspectives of the participants rather than the researcher.
Repeated periods of fieldwork have allowed model strategy, parameters and assump-
tions to be tested and refined with the original research participants.