Representing Human Decision-Making in Environmental Modelling - Environmental Modelling: Finding Simplicity in Complexity - page 282

If you are still with us, then hopefully you agree that some
attempt to represent human decision-making in environ-
ment modelling is necessary - especially if the models we
come up with help us understand why we make certain
decisions (both rational and irrational) and what this
means for the environment. In this chapter we outline
several approaches for the representation and integration
of human activity and decision-making in environmental
modelling. We present these approaches in three broad
categories: Scenarios and Integrated Assessment Mod-
elling, Economic Modelling, and Agent-Based Modelling
(but remain conscious of overlap between the meth-
ods in the categories). These approaches vary in how
well they represent reciprocal interactions (e.g. scenario
approaches poorly, agent-based approaches potentially
very well), their representation of heterogeneity in indi-
viduals' decision-making processes (traditional economic
models poorly, agent-based models potentially well), and
the tractability of the influence of model assumptions
on outcomes (agent-based approaches potentially low,
scenario approaches high). We discuss some of the wider
issues of modelling humans and highlight some of the
key issues that modellers will need to consider. The
examples we present are only a brief introduction to this
wide-ranging modelling activity and we encourage those
seeking more detail to refer to the cited references as
starting points for further reading.

construction of the flood defences? What if more homes
are built on floodplains as developers see opportunities
in apparently safe areas, without accounting for the ways
in which this urbanization will affect the hydrological
regime? This urbanization might produce more rapid
runoff into the system and lower the threshold rainfall
required to cause overtopping of the defences. Developers
may make such decisions because they know that the costs
of damage will be met through government assistance and
insurance claims and therefore spread across the popula-
tion and thus be bearable (although recent developments
suggest that this may not be for much longer). What if
land use upstream changes? Whether the answer is more
urbanization or a change in crop cycles (see Boardman,
2001, on changes relating to winter-wheat production in
southern England), the production of runoff from the
land surface may change again. Such land-use changes
may even be encouraged as people believe the downstream
sites can cope due to the presence of the new defences.
Considering these questions using the simulation model
before defences are constructed will give a more realis-
tic simulation of the effect of the intervention and may
prevent future costs or even disaster. Even at such local
scales, such 'what if?' questions are by no means simple
to answer.
At greater spatial and temporal extents scenarios can
become vastly more complex and numerous. The general
circulation model (GCM) simulations conducted under
the auspices of the Intergovernmental Panel on Climate
Change (IPCC) are possibly the best known scenario
approach at the global scale. Continued scenario devel-
opment since the IPCC Special Report on Emissions
Scenarios (SRES; IPCC, 2000) means there are now more
than 800 greenhouse-gas (GHG) emissions scenarios in
the literature (Nakicenovic et al ., 2006), many developed
with the aid of integrated assessment models (IAMs, see
below). In turn, these scenarios (themselves the out-
comes of models) form the inputs for the GCMs that
produce a set of climate impacts (which in turn are often
applied as inputs to models of hydrology, crop growth or
other systems analysed by the so-called 'climate impacts'
community of modellers).
The SRES scenarios provide a good example of how
elaborate the development of scenarios can become. The
many newer scenarios suggest lower mean global popu-
lation and lower maximum primary energy production
than the SRES scenarios but differences are minimal and
the SRES scenarios were used again in the fourth IPCC
assessment report (Nakicenovic et al ., 2006; Fisher et al .,
2007). The SRES scenarios were based on four qualitative

18.2 Scenario approaches

Scenarios are a common way of representing the effects
of human decision making on the environment. They can
be used at scales ranging from local to global. Scenarios
are neither predictions nor forecasts but allow us to ask
'what if ... ?' questions. In themselves they are a form of
modelling and range from simple thought experiments of
the type used by large corporations such as Shell to map
the future and how they should prepare for it, to elaborate
systems models built as part of research projects. The pri-
mary use of scenarios in modelling human-environment
interactions has been as a tool to represent potential
human activity as an input to models for the exploration
of alternative environmental system states and futures.
As an example, consider a simulation model that eval-
uates the hydrological response to the construction of
flood defences in particular low-lying areas. The model
may even account for predicted future changes in climate
and sea-level rise. But what about all the corresponding
future decisions that may be taken as a response to the