System Dynamics Modelling
Introduction
System dynamics was developed for the purpose of characterizing complex, non-linear systems through capturing interrelations, feedback loops and delays (Langsdale et al, 2007). System dynamics models (SDMs) are liked because of their simplicity – they are summarising models that relate aggregate variables (populations, sectors, nations etc.). Because of the visualization of the model provided by the graphical construction interface they are ideal for use with participant groups (eg. to facilitate dialogue or to elicit observations about causal relations). These models can easily manage both clearly- defined and poorly-defined components in the same model. Similarly, they can capture quantitative, physical parts of the system, such as hydrology, as well as intangible parts of the system, such as policies and human responses, and they can be used to explore different types of sensitivities and to compare outcomes of different scenarios. Models used for collaborative modeling in water resources applications include system dynamics platforms like STELLA (Cardwell et al., 2004; Costanza & Ruth, 1998; Langsdale et al., 2006), Vensim and Studio Expert (Tidwell et al., 2004) whilst it is also possible to use more general-purpose computing software such as MATLAB. Other types of models which have been used include MIKE-BASIN (Borden & Spinazola, 2006; Borden et al., 2006); the Water Evaluation And Planning system model (WEAP) (Jenkins et al., 2005). System dynamics software packages are blank slates and can be applied to any problem, while MIKE-BASIN, WEAP, and OASIS are all limited to water resources applications. Case studies where the system dynamics approach was applied to environmental issues include: water resources management in Switzerland, Senegal and Thailand, and vegetation management in Zimbabwe (Hare et al., 2003); water allocation issues in the Namoi River, Australia (Letcher & Jakeman, 2003). Perhaps the most famous application of system dynamics modelling was done as part of the Club of Rome report named ‘Limits to Growth’ (Meadows et al. 1972) which illustrated the possible population, resource and pollution catastrophe facing the planet based on hypothesised relationships \ feedbacks among the key macro variables. It also illustrates the sensitivity, often inherent in simulation models, of results to small changes in parameter values. In the case of ‘Limits to Growth’, where it was hard to empirically justify \ confirm parameter settings of the model, this lead to criticism over the conclusions drawn by the authors and the claimed predictive ability of the model. This influential, landmark study, revealed the difficulty of performing reliable long-term analyses – notwithstanding the model used was rudimentary compared to what can be constructed using today’s software platforms. Social simulation approaches now also include agent-based modelling (ABM) – which contrasts a concentration on individual actors and their interrelations with SDMs focus on aggregate relations. See also the research article by HJ Scholl (2001) comparing the two approaches.ABM allows detail to be included at the lower level of analysis (e.g the distributed nature of some facet of the model) with simulations being used to explore macro-scale impacts of this ‘complication’. However there are also many ‘hybrid’ multi- level applications linking micro ABMs to aggregate SDMs, as discussed below.
Decision-making applications of SDMs
There are, however some examples of the use of a systems modelling in adaptation decision making. The WEAP model was adapted to investigate the Tana river basin in Kenya (Dyszynski et al., 2009), where various socio-economic scenarios were used with projections for the period 2045-2065 in minimum temperature, maximum temperature and precipitation for the station of Meru. Projections were extracted from Global Circulation models. Impacts of climate change scenarios on water availability, irrigation water shortage, urban water shortage and hydropower generation and costs of impacts were calculated both with and without a range of adaptation options, both supply and demand side.
In the EU funded Newater project, Hetman et al. (2006) investigated the role of Agent Based Modelling vs. System Dynamics in the dynamics of farming in river basins asking whether models provide insight on the future of farming in rural areas. The two main purposes of the research were: to compare two modeling approaches and to examine the effect of different structures of social interactions on the behaviour of farmers. These two modeling approaches are: an integrated model joining the components representing a network of interacting agents (social part) and the aggregated components representing the dynamics of an ecological system (a hybrid, e.g. a linked agent-based and system dynamics model) and a pure system dynamics model.
A toolkit concept to support policy-making and planning based on a Global Environment and National Information Evaluation System (GENIES, Shahet al, 2012) is being developed that focuses on the core issues of adaptation, mitigation, risk, and economics of climate change and how they interrelate with aspects of water, energy, the built environment, transport, waste, and ecosystems. While recognizing the plethora of methodological perspectives that pertain to each sector, a system dynamics method is used, which lends itself to integrated assessment, given its flexibility and ease of extension and revision as new policy and planning questions emerge. The framework design starts with a clear definition of a problem and then draws together the appropriate models and data, to enable relationships to be defined and processed in a scientifically robust manner to evaluate adaptation and mitigation options.
In a review of the use of system dynamics in decision making Pruyt (2006) concludes that stand-alone use of system dynamics for decision-making in case of such issues also shows several shortcomings. First, stand-alone use of system dynamics does not automatically answer the questions as to what views, dimensions and time frames are of most importance and should be included. Secondly, there use does not take multiple (conflicting) views and multiple dimensions over time into account. It therefore can not indicate how to select a strategy when no strategy is overall the ‘best’ and most robust at any time on all dimensions for all parties given all uncertainties. Thirdly, it does not allow the consideration/integration in the strategy selection of many other types of information (mostly qualitative) concerning robustness, uncertainties, etc. Stand-alone use of system dynamics is considered to be less appropriate for evaluation and choice when the strategy selection is not obvious or is ambiguous.
References
Borden, C. & Spinazola, B. G. J. (2006), Evaluation of diversions operation plans to meet minimum fish flow requirements using MIKE BASIN model simulations in Lehmi River Basin, in ‘Adaptive Management of Water Resources’, American Water Resources Association, Missoula, Montana. 56
Borden, C., Wilfert, J., Chester, J. & Tsang, M. (2006), Water supply assessment tool for the San Fransisco Public Utility Commission, in ‘Adaptive Management of Water Resources’, American Water Resources Association, Missoula, Montana. 56
Cardwell, H., Faber, B. & Spading, K. (2004), Collaborative models for planning in the Missisippi headwaters, in ‘World Water and Environmental Resources Congress’, American Society of Civil Engineers, Salt Lake City, Utah. 56.
Costanza, R. & Ruth, M. (1998), ‘Using dynamic modeling to scope environmental problems and build consensus’, Environmental Management 22(2), 183– 195. 56, 57
Dyszynski, J., Droogers, P. and Butterfield, R., 2009. Climate adaptation economics. Kenya water sector. Report for DFID Economics of Climate Change in Kenya project. Stockholm Environment Institute. 2009.
Hare, M., Letcher, R. A. & Jakeman, A. J. (2003), ‘Participatory modelling in natural resources management: A comparison of four case studies’, Integrated Assessment 4(2), 62–72. 5
Jenkins, M., Marques, G. & Lelo, F. (2005), WEAP as a participatory tool for shared vision planning in the River Njoro Watershed, Kenya, in ‘Impacts of global climate change: Proceedings of the 2005 World Water & Environmental Resources Congress’, ASCE, Anchorage, Alaska. 56
Langsdale, S., Beall, A., Carmichael, J., Cohen, S. & Forster, C. (2006), Chapter 5: Exploring water resources futures with a system dynamics model, in S. C hen & T. Neale, eds, ‘Participatory integrated assessment of water management and climate change in the Okanagan Basin, British Columbia, Canada: Final Report.’, University of British Columbia and Environment Canada, Vancouver, BC. 55, 56, 62.
Langsdale, S., Beall, A., Carmichael, J., Cohen, S. & Forster, C. , 2007. An Exploration of Water Resources Futures under Climate Change Using System Dynamics Modeling. The Integrated Assessment Journal. Bridging Sciences & Policy, Vol. 7, Iss. 1 (2007), Pp. 51–79.
Letcher, R. A. & Jakeman, A. J. (2003), ‘Application of an adaptive method for integrated assessment of water allocation issues in the Namoi River catch- ment, Australia’, Integrated Assessment 4(2), 73–89. 57
Meadows DH, Meadows D, Randers J, Behrens WWIII (1972) The Limits to Growth: a report for the Club of Rome’s project on the predicament of mankind. Universe Books, New York
Pruyt, E. 2006. System Dynamics and Decision-Making in the Context of Dynamically Complex Multi-Dimensional Societal Issues. System Dynamics Conference.
Scholl HJ. 2001. Agent-based and system dynamics modeling : A call for cross study and joint research. Presented at 34th Hawaii International Conference on System Sciences – 2001.
Shah, J., Urich, P., Li, Y., Ye, W., Carr, R. (2012). Global Environment and National Information Evaluation System (GENIES) for Urban Impact Analysis. Greater Mekong Subregion Conference 2020, Bangkok, Thailand.
Context
It has become clear that the most common techniques used in appraisal (and decision support) have limitations in coping with the uncertainty associated with the climate change decision context (e.g. see Hunt and Watkiss, 2011). As a result, there is a growing consensus that economic appraisal of climate change adaptation should incorporate the multiple sources of uncertainty.
While the focus on decision making under uncertainty has become widespread in the adaptation literature, with a focus on iterative adaptation management (Adger et al, 2006; Downing et al, 2012), there has been much less adoption of these concepts in the domain of economic assessment and appraisal. Indeed, the default position is to rely on conventional approaches, notably cost-benefit analysis, which has traditionally been adopted in appraisal of sectoral development plans and projects. Further, while it is recognised that new economic decision support tools may help account for uncertainty, there are relatively few practical adaptation applications to date.
Against this background, IMPACT2C project has critically reviewed a number of approaches for potential use in adaptation appraisal, and search for examples of application. This includes system dynamics modelling.
Related pages
Water Evaluation and Planning System (WEAP)
Social network analysis/mapping
See more Adaptation decision-making methods
Related projects
IMPACT2C Quantifying climate impacts, vulnerabilities, risks and economic costs, as well as potential adaptive responses, at pan-European scale